[House Hearing, 116 Congress]
[From the U.S. Government Publishing Office]
THE FUTURE OF WORK: PROTECTING
WORKERS' CIVIL RIGHTS IN THE
DIGITAL AGE
=======================================================================
HEARING
before the
SUBCOMMITTEE ON CIVIL RIGHTS AND HUMAN SERVICES
COMMITTEE ON EDUCATION
AND LABOR
U.S. HOUSE OF REPRESENTATIVES
ONE HUNDRED SIXTEENTH CONGRESS
SECOND SESSION
__________
HEARING HELD IN WASHINGTON, DC, FEBRUARY 5, 2020
__________
Serial No. 116-50
__________
Printed for the use of the Committee on Education and Labor
[GRAPHIC(S) NOT AVAILABLE IN TIFF FORMAT]
Available via the: https://edlabor.house.gov or www.govinfo.gov
______
U.S. GOVERNMENT PUBLISHING OFFICE
39-731 PDF WASHINGTON : 2021
COMMITTEE ON EDUCATION AND LABOR
ROBERT C. ``BOBBY'' SCOTT, Virginia, Chairman
Susan A. Davis, California Virginia Foxx, North Carolina,
Raul M. Grijalva, Arizona Ranking Member
Joe Courtney, Connecticut David P. Roe, Tennessee
Marcia L. Fudge, Ohio Glenn Thompson, Pennsylvania
Gregorio Kilili Camacho Sablan, Tim Walberg, Michigan
Northern Mariana Islands Brett Guthrie, Kentucky
Frederica S. Wilson, Florida Bradley Byrne, Alabama
Suzanne Bonamici, Oregon Glenn Grothman, Wisconsin
Mark Takano, California Elise M. Stefanik, New York
Alma S. Adams, North Carolina Rick W. Allen, Georgia
Mark DeSaulnier, California Lloyd Smucker, Pennsylvania
Donald Norcross, New Jersey Jim Banks, Indiana
Pramila Jayapal, Washington Mark Walker, North Carolina
Joseph D. Morelle, New York James Comer, Kentucky
Susan Wild, Pennsylvania Ben Cline, Virginia
Josh Harder, California Russ Fulcher, Idaho
Lucy McBath, Georgia Steve Watkins, Kansas
Kim Schrier, Washington Ron Wright, Texas
Lauren Underwood, Illinois Daniel Meuser, Pennsylvania
Jahana Hayes, Connecticut Dusty Johnson, South Dakota
Donna E. Shalala, Florida Fred Keller, Pennsylvania
Andy Levin, Michigan* Gregory F. Murphy, North Carolina
Ilhan Omar, Minnesota Jefferson Van Drew, New Jersey
David J. Trone, Maryland
Haley M. Stevens, Michigan
Susie Lee, Nevada
Lori Trahan, Massachusetts
Joaquin Castro, Texas
* Vice-Chair
Veronique Pluviose, Staff Director
Brandon Renz, Minority Staff Director
------
SUBCOMMITTEE ON CIVIL RIGHTS AND HUMAN SERVICES
SUZANNE BONAMICI, OREGON, Chairwoman
Raul M. Grijalva, Arizona James Comer, Kentucky,
Marcia L. Fudge, Ohio Ranking Member
Kim Schrier, Washington Glenn ``GT'' Thompson,
Jahana Hayes, Connecticut Pennsylvania
David Trone, Maryland Elise M. Stefanik, New York
Susie Lee, Nevada Dusty Johnson, South Dakota
C O N T E N T S
----------
Page
Hearing held on February 5, 2020................................. 1
Statement of Members:
Bonamici, Hon. Suzanne, Chairwoman, Subcommittee on Civil
Rights and Human Services.................................. 1
Prepared statement of.................................... 4
Comer, Hon. James, Ranking Member, Subcommittee on Civil
Rights and Human Services.................................. 5
Prepared statement of.................................... 6
Statement of Witnesses:
Ajunwa, Ms. Ifeoma, J.D., Ph.D. Assistant Professor of
Employment and Labor Law, Cornell University............... 13
Prepared statement of.................................... 15
Romer-Friedman, Mr. Peter, J.D., Principal and Head of the
Civil Rights and Class Actions Practice Gupta Wessler PLLC. 26
Prepared statement of.................................... 28
Lander, Ms. Esther G., J.D., Partner, Akin Gump Strauss Hauer
and Feld LLP............................................... 16
Prepared statement of.................................... 18
Yang, Ms. Jenny R., Senior Fellow, Urban Institute........... 9
Prepared statement of.................................... 11
Additional Submissions:
Chairwoman Bonamici:.........................................
Letter dated February 3, 2020 from The Leadership
Conference on Civil and Human Rights................... 50
Link: Help Wanted: An Examination of Hiring Algorithms,
Equity, and Bias....................................... 52
Mr. Comer:...................................................
Prepared statement from the HR Policy Association........ 53
Questions submitted for the record by:
Chairwoman Banamici
Hayes, Hon. Jahana, a Representative in Congress from the
State of Connecticut................................... 60
Responses submitted for the record by:
Mr. Romer-Friedman....................................... 63
Ms. Yang................................................. 69
THE FUTURE OF WORK: PROTECTING
WORKERS' CIVIL RIGHTS IN THE
DIGITAL AGE
----------
Wednesday, February 5, 2020
House of Representatives,
Subcommittee on Civil Rights and Human Services,
Committee on Education and Labor
Washington, DC
----------
The subcommittee met, pursuant to call, at 2:03 p.m., in
Room 2175, Rayburn House Office Building. Hon. Suzanne Bonamici
[chairwoman of the subcommittee] presiding.
Present: Representatives Bonamici, Schrier, Lee, Comer,
Stefanik, and Johnson.
Also present: Representatives Scott, Foxx, Takano, and
Blunt Rochester.
Staff present: Tylease Alli, Chief Clerk; Ilana Brunner,
General Counsel; Emma Eatman, Press Assistant; Eunice Ikene,
Labor Policy Advisor; Stephanie Lalle, Deputy Communications
Director; Andre Lindsay, Staff Assistant; Jaria Martin, Clerk/
Special Assistant to the Staff Director; Kevin McDermott,
Senior Labor Policy Advisor; Richard Miller, Director of Labor
Policy; Max Moore, Staff Assistant; Veronique Pluviose, Staff
Director; Carolyn Ronis, Civil Rights Counsel; Banyon Vassar,
Deputy Director of Information Technology; Katelyn Walker,
Counsel; Rachel West, Senior Economic Policy Advisor; Gabriel
Bisson, Minority Staff Assistant; Courtney Butcher, Minority
Director of Member Services and Coalitions; Rob Green, Minority
Director of Workforce Policy; Jeanne Kuehl, Minority
Legislative Assistant; John Martin, Minority Workforce Policy
Counsel; Hannah Matesic, Minority Director of Operations;
Carlton Norwood, Minority Press Secretary; and Ben Ridder,
Minority Professional Staff Member.
Chairwoman BONAMICI. The Committee on Education and Labor
will come to order. Welcome, everyone. I note that a quorum is
present.
The Committee is meeting today for a legislative hearing to
hear testimony on ``The Future of Work, Protecting Workers'
Civil Rights in the Digital Age.''
I note for the Subcommittee that Congressman Mark Takano of
California, Congresswoman Pramila Jayapal of Washington,
Congresswoman Lori Trahan of Massachusetts, Congresswoman
Yvette Clark of New York, and Congresswoman Lisa Blunt
Rochester of Delaware will be permitted to participate in
today's hearing with the understanding that their questions
will come only after all Members of this Subcommittee and then
the Full Committee on both sides of the aisle who are present
have had an opportunity to question the witnesses.
I will now move to opening statements. Pursuant to
Committee Rule 7(c), opening statements are limited to the
Chair and the Ranking Member. This allows us to hear from our
witnesses sooner and provides all Members with adequate time to
ask questions. I now recognize myself for the purpose of an
opening statement.
Technology and automation have become entrenched in nearly
every aspect of our society and culture. The intentions behind
the use of technology may be noble, but our efforts to both
assess and address the effects on our workplace have been
inadequate.
In recent years, employers have harnessed new digital tools
like recruiting and hiring algorithms, computer analyzed video
interviews, and real time tracking of their workers, to cut the
cost of hiring and managing workers.
This is our third hearing in our Future of Work series and
today we will examine how the technologies that employers use
for hiring and management may, intentionally or not, facilitate
discrimination and undermine workers' civil rights. We will
discuss how Congress, Federal agencies, and the business
community can strengthen workplace protections to make sure
workers are not left vulnerable to discriminatory practices.
And to prevent discriminatory hiring, firing, and
monitoring practices, we will investigate whether new
technologies are designed to account for implicit and explicit
bias and are used transparently.
Proponents of new technologies assert that digital tools
eliminate bias and discrimination by attempting to remove
humans from the process. But technology is not developed or
used in a vacuum.
A growing body of evidence suggests that, left unchecked,
digital tools can absorb and replicate systemic biases that are
ingrained in the environment in which they are designed.
For example, hiring algorithms often rely on correlations
to make predictions about the capabilities of job candidates.
Yet these tools can mistake correlation for causation and
subsequently perpetrate harmful disparities.
In 2017, an algorithm built by Amazon to hire engineers was
scrapped after it was found to favor men over women by
penalizing graduates of women's colleges. Because men hold the
majority of engineering positions, the algorithm had presumed
that being male was a key characteristic of successful
engineers when in reality, being male does not cause one to be
a successful engineer.
New technologies that surveil and monitor workers can also
exacerbate bias in the workplace. These tools may force workers
to share their location, activities, and even private biometric
information, sometimes without workers' knowledge or consent.
The technologies also allow employers to access private
information that could be used to discriminate against workers.
For instance, through certain workplace wellness programs, an
employer could learn of a disability, a health condition, or
genetic condition that is otherwise protected by
antidiscrimination law.
Too often employers and technology vendors are not
transparent about the design and use of digital tools, posing
challenges for workers seeking redress for workplace
discrimination.
Simply put, without transparent and responsible design,
digital tools can further perpetuate and even exacerbate long-
held biases that have led to workplace disparities,
particularly for women of colors--color, individuals, women--
individuals with disabilities, women, and older workers.
Moreover, digital tools that are opaque in their design and
operation cannot be held accountable. As traditional employment
relationships shift dramatically in our modern economy,
workers' antidiscrimination protections are also in jeopardy.
As this Committee has established, new technologies have
fundamentally restructured the workplace through the rise of
gig and platform work.
These platforms have provided workers with new
opportunities, but many employers have also used new
technologies to deny workers basic protections.
For example, app-based companies frequently misclassify
their employees as independent contractors, depriving them of
protections and benefits such as minimum wage and overtime pay.
Worker misclassification is not unique to app-based
companies. Some app-based companies directly hire their
employees, as we learned from a business leader in our first
Future of Work hearing.
Workers misclassified as independent contractors are also
excluded from the majority of Federal workplace
antidiscrimination laws, including protection under Title VII
of the Civil Rights Act of 1964, the Americans with
Disabilities Act, and the Age Discrimination in Employment Act.
These gaps leave workers classified as independent
contractors, whether misclassified or not, with few options to
challenge discrimination.
We have the responsibility on this Committee to work with
Federal agencies and the business community to strengthen
workplace protections in the face of changing technology. And
this should include the right to be free from workplace
discrimination and the right to be hired based on
qualifications rather than age, identity, or zip code.
We must compel employers and technology vendors to be
transparent and accountable for new workplace technologies. We
must invest in our key defenses against employment
discrimination and empower the Equal Employment Opportunity
Commission to address emerging forms of digital discrimination
and we must identify and close the gaps in our Nation's laws
that leave workers vulnerable to misclassification,
discrimination, and harassment on the job.
I request unanimous consent to enter into the record a
letter from The Leadership Conference on Civil and Human Rights
and Upturn and a recent report on hiring algorithms, equity,
and bias from Upturn into the record.
Without objection, so ordered.
I look forward to our discussion today, and I now yield to
the Ranking Member, Mr. Comer, for an opening statement and I
do want to note, I went long so if you want to take a little
extra time, feel free.
[The statement of Chairwoman Bonamici follows:]
Prepared Statement of Hon. Suzanne Bonamici, Chairwoman, Subcommittee
on Civil Rights and Human Services
Technology and automation have become entrenched in nearly every
aspect of our society and culture. The intentions behind the use of
technology may be noble, but our efforts to both assess and address the
effects on our workforce have been inadequate. In recent years,
employers have harnessed new digital tools--like recruiting and hiring
algorithms, computer-analyzed video interviews, and real-time tracking
of their workers--to cut the cost of hiring and managing workers.
This is our third hearing in our Future of Work series. Today we
will examine how the technologies that employers use for hiring and
management may, intentionally or not, facilitate discrimination and
undermine workers' civil rights. We will discuss how Congress, federal
agencies, and the business community can strengthen workplace
protections to make sure workers are not left vulnerable to
discriminatory practices. And, to prevent discriminatory hiring,
firing, and monitoring practices, we will investigate whether new
technologies are designed to account for implicit and explicit bias and
are used transparently.
Proponents of new technologies assert that digital tools eliminate
bias and discrimination by attempting to remove humans from the
processes. But technology is not developed or used in a vacuum. A
growing body of evidence suggests that, left unchecked, digital tools
can absorb and replicate systemic biases that are ingrained in the
environment in which they are designed.
For example, hiring algorithms often rely on correlations to make
predictions about the capabilities of job candidates. Yet these tools
can mistake correlation for causation and subsequently perpetuate
harmful disparities. In 2017, an algorithm built by Amazon to hire
engineers was scrapped after it was found to favor men over women by
penalizing graduates of women's colleges. Because men hold the majority
of engineering positions, the algorithm had presumed that being male
was a key characteristic of successful engineers. In reality, being
male does not cause one to be a successful engineer.
New technologies that surveil and monitor workers can also
exacerbate bias in the workplace. These tools may force workers to
share their location, activities, and even private biometric
information--sometimes without workers' knowledge or consent. The
technologies also allow employers to access private information that
could be used to discriminate against workers. For instance, through
certain workplace wellness programs, an employer could learn of a
disability, health condition, or genetic condition that is otherwise
protected by anti-discrimination law.Too often employers and technology
vendors are not transparent about the design and use of digital tools,
posing challenges for workers seeking redress for workplace
discrimination.
Simply put, without transparent and responsible design, digital
tools can further perpetuate and even exacerbate long-held biases that
have led to workplace disparities, particularly for workers of color,
women, individuals with disabilities, and older workers. Moreover,
digitial tools that are opaque in their design and operation cannot be
held accountable.
As traditional employment relationships shift dramatically in our
modern economy, workers' antidiscrimination protections are also in
jeopardy. As this Committee has established, new technologies have
fundamentally restructured the workplace through the rise of ``gig''
and ``platform'' work. These platforms have provided workers with new
opportunities, but many employers have also used new technologies to
deny workers basic protections.
For example, app-based companies frequently misclassify their
employees as ``independent contractors,'' depriving them of protections
and benefits such as minimum wage and overtime pay. Worker
misclassification is not unique to app-based companies. Some app-based
companies directly hire their employees, as we learned from a business
leader in our first Future of Work hearing.
Workers misclassified as independent contractors are also excluded
from the majority of federal workplace antidiscrimination laws,
including protections under Title VII of the Civil Rights Act of 1964,
the Americans with Disabilities Act, and the Age Discrimination in
Employment Act. These gaps leave workers classified as independent
contractors--whether misclassified or not--with few options to
challenge workplace discrimination.
We have the responsibility on this Committee to work with federal
agencies and the business community to strengthen workplace protections
in the face of changing technology. And this should include the right
to be free from workplace discrimination and the right to be hired
based qualifications rather than age, identity, or zip code.
We must compel employers and technology vendors to be transparent
and accountable for new workplace technologies. We must invest in our
key defenses against employment discrimination, and empower the Equal
Employment Opportunity Commission to address emerging forms of digital
discrimination. And we must identify and close the gaps in our nation's
laws that leave workers vulnerable to misclassification,
discrimination, and harassment on the job.
I request unanimous consent to enter a letter from The Leadership
Conference on Civil and Human Rights and Upturn and a recent report on
hiring algorithms, equity, and bias from Upturn into the record.
I look forward to our discussion today, and I now yield to the
Ranking Member, Mr. Comer, for an opening statement.
______
Mr. COMER. All right. Well, thank you, Madam Chair, and
today we are here to discuss how technological advancements are
affecting workers.
New technologies continue to increase efficiency, reduce
costs for employers in recruiting and hiring and lead to
quicker job placements and enhanced job opportunities.
In a statement to this Committee, the HR Policy Association
noted, quote, in a recent survey 71 percent of staffing firms
believe artificial intelligence will eliminate human bias from
the recruitment process, unquote.
So not only can employers utilize new technologies to
eliminate employment bias, but they can also be used to
decrease time and the cost of doing business.
Technology has also driven the sharing economy which has
created substantial opportunities for workers and job creators
who are seeking flexible workforce arrangements so they can
better compete in our ever-changing economy.
Workers are seeking out the benefits and flexibility these
arrangements provide as they recognize how significantly they
can improve their quality of life as well as their family's.
This is a growing trend among American workers and job seekers
that should be encouraged, not impeded.
Many businesses who also value flexibility and productivity
are turning to independent contractors. The use of independent
contractors makes sense for job creators looking to obtain
high-quality services, for workers who want to offer their
skills on their own terms, and for consumers who benefit from a
reduction in the cost of goods and services.
Simply put, online platforms and other emerging
technologies have given American workers more control,
flexibility, and opportunity in the workplace than they have
previously had. Regardless of technological advancements, every
American should have the opportunity to achieve success in the
workplace free from discrimination.
This is why there are important protections built into
Federal law to prevent workplace discrimination. These
protections are broadly written and continue to apply to new
and emerging technologies.
These laws protect individuals from employment,
discrimination based on age, color, disability, genetic
information, national origin, race, religion, or sex.
Workers in the sharing economy are also protected. For
example, the Fair Labor Standards Act has strong remedies in
place for employers who incorrectly classify workers and
violate minimum wage and overtime requirements.
All workers should be paid in full for their work. That is
why Committee Republicans support enforcement of the FLSA. We
shouldn't penalize Americans who work for themselves or the
companies that do businesses with them.
Instead, we should applaud these Americans for their
entrepreneurial spirit. Our Nation's laws were written so that
they can be and are applied to employers' use of technologies
in ways that protect workers.
Additionally, it should go without saying that the
overwhelming majority of businesses follow the law and want to
do what is expected of them. Bottom line, workers, job
creators, and the U.S. economy are all benefitting from today's
technological advancements.
Madam Chair, before we hear from our witnesses, I need to
take a moment to point out the hypocrisy of today's hearing. My
Democrat colleagues want to talk about protecting workers'
rights while they simultaneously push radical legislation that
will undermine the rights of workers.
H.R. 2474, the PRO Act, which we expect will be on the
House floor for a vote tomorrow, is written to bail out the
failing labor union business model that is being widely
rejected by American workers.
This radical legislation would penalize entrepreneurships
by creating an expansive, one-size-fits-all definition of an
employee, which will increase costs for business owners as well
as consumers while limiting worker opportunities for
individuals who desire flexibility.
Instead, we should champion reforms that expand
opportunities for flexibility, innovation, and entrepreneurship
to give workers and job seekers opportunities to compete
successfully in the modern economy.
I thank the witnesses for being here today and I look
forward to their testimony and, Madam Chair, I yield back.
[The statement of Mr. Comer follows:]
Prepared Statement of Hon. James Comer, Ranking Member, Subcommittee on
Civil Rights and Human Services
``Today, we are here to discuss how technological advancements are
impacting workers.
New technologies continue to increase efficiency, reduce costs for
employers in recruiting and hiring, and lead to quicker job placements
and enhanced job opportunities. In a statement to this Committee, the
HR Policy Association noted: `In a recent survey, 71 percent of
staffing firms believe artificial intelligence will eliminate human
bias from the recruitment process.' So, not only can employers utilize
new technologies to eliminate employment bias, but they can also be
used to decrease the time and cost of doing business.
Technology has also driven the sharing economy, which has created
substantial opportunities for workers and job creators who are seeking
flexible workforce arrangements so they can better compete in our ever-
changing economy. Workers are seeking out the benefits and flexibility
these arrangements provide as they recognize how significantly they can
improve their quality of life, as well as their families. This is a
growing trend among American workers and jobseekers that should be
encouraged, not impeded. Many businesses who also value flexibility and
productivity are turning to independent contractors. The use of
independent contractors makes sense for job creators looking to obtain
high-quality services, for workers who want to offer their skills on
their own terms, and for consumers who benefit from a reduction in the
cost of goods and services.
Simply put, online platforms and other emerging technologies have
given American workers more control, flexibility, and opportunity in
the workplace than they have previously had.
Regardless of technological advancements, every American should
have the opportunity to achieve success in the workplace free from
discrimination. That is why there are important protections built into
federal law to prevent workplace discrimination. These protections are
broadly written and continue to apply to new and emerging technologies.
These laws protect individuals from employment discrimination based
on age, color, disability, genetic information, national origin, race,
religion, or sex.
Workers in the sharing economy are also protected. For example, the
Fair Labor Standards Act (FLSA) has strong remedies in place for
employers who incorrectly classify workers and violate minimum wage and
overtime requirements. All workers should be paid in full for their
work. That is why Committee Republicans support enforcement of the
FLSA. We shouldn't penalize Americans who work for themselves or the
companies that do business with them. Instead, we should applaud these
Americans for their entrepreneurial spirit.
Our nation's laws were written so that they can be, and are,
applied to employers' use of technologies in ways that protect workers.
Additionally, it should go without saying that the overwhelming
majority of businesses follow the law and want to do what is expected
of them. Bottom line, workers, job creators, and the U.S. economy are
all benefiting from today's technological advancements.
Madam Chair, before we hear from our witnesses, I need to take a
moment to point out the hypocrisy of today's hearing. My Democrat
colleagues want to talk about protecting workers' rights while they
simultaneously push radical legislation that will undermine the rights
of workers.
H.R. 2474, the PRO Act, which we expect will be on the House floor
for a vote tomorrow, is written to bail out the failing labor union
business model that is being widely rejected by American workers. This
radical legislation would penalize entrepreneurship by creating an
expansive, one-size-fits-all definition of an employee, which will
increase costs for business owners as well as consumers, while limiting
work opportunities for individuals who desire flexibility.
Instead, we should champion reforms that expand opportunities for
flexibility, innovation, and entrepreneurship to give workers and job
seekers opportunities to compete successfully in the modern economy.
I thank the witnesses for being here and I look forward to their
testimony.''
______
Chairwoman BONAMICI. Thank you, Mr. Comer. I know we will
be having the PRO Act debate on the floor as well as in this
Committee but now we are going to focus on the topic at hand.
Without objection, all other Members who wish to insert written
statements into the record may do so by submitting them to the
Committee Clerk electronically in Microsoft Word format by 5
p.m. on Tuesday, February 18, 2020.
I will now introduce our distinguished panel of witnesses
and I will introduce each witness before we begin questions.
First, Ms. Jenny Yang served as the Chair of the U.S. Equal
Employment Opportunity Commission from September of 2014 to
January of 2017, and as Vice Chair and a Member of the
Commission from 2013 to 2018.
Under her leadership, the commission launched the Select
Task Force on the Study of Harassment in the Workplace to
identify innovative solutions to prevent harassment at work.
And she led efforts to strengthen the EEOC's annual data
collection to include employer reporting of pay data.
Next, we have Dr. Ifeoma Ajunwa. She is an assistant
professor of labor and employment law in the Law, Labor
Relations, and History Department of Cornell University's
Industrial and Labor Relations School and an associate faculty
member at Cornell Law School.
She is also a faculty associate at the Berkman Kline Center
at Harvard Law School and an affiliate of The Center for the
Study of Inequality at Cornell University. She is a 2019
recipient of the National Science Foundation Career Award and a
2018 recipient of the Derrick A. Bell award from the
Association of American Law Schools.
Dr. Ajunwa's research interests are at the intersection of
law and technology with a particular focus on the ethical
governance of workplace technologies.
And at the discretion of the Chair, I do want to mention
that Derrick Bell was my law school dean when I went to law
school at the University of Oregon, so it is an honor that you
are here with that award, that distinguished award.
Ms. Esther Lander is a partner at Akin Gump Strauss Hauer &
Feld LLP, Washington, D.C., where she focuses on complex
employment litigation, high-stakes internal and government
investigations, and client counseling.
She previously served as the Principal Deputy Chief of the
Employment Litigation Section within the Civil Rights Division
at the Department of Justice.
Mr. Peter Romer-Friedman is a principal at Gupta Wessler
PLLC in Washington, D.C., where he heads the firm's new civil
rights and class actions practice.
He maintains a dynamic and innovative civil rights docket
with an emphasis on employment discrimination and benefits,
fair housing, credit discrimination, and constitutional rights.
The civil rights cases often arise at the cutting edge of the
law and focus on solving both entrenched and emerging problems
with novel approaches.
We appreciate all of the witnesses for being here today and
we look forward to your testimony. Let me remind the witnesses
that we have read your written statements and they will appear
in full in the hearing record.
Pursuant to Committee Rule 7(d) and Committee practice,
each of you is asked to limit your oral presentation to a 5
minute summary of your written statement.
Let me remind the witnesses as well that pursuant to Title
18 of U.S. Code Section 1001, it is illegal to knowingly and
willfully falsify any statement, representation, writing,
document, or material fact presented to Congress or otherwise
conceal or cover up a material fact.
Before you begin your testimony, please remember to press
the button on your microphone in front of you so it will turn
on and the Members can hear you.
As you begin to speak, the light in front of you will turn
green. After 4 minutes, the light will turn yellow to signal
that you have 1 minute remaining. When the light turns red,
your 5 minutes have expired and we ask you to wrap up. We will
let the entire panel make their presentations before we move to
Member questions. When answering a question, again, please
remember to turn your microphone on, and I first recognize Ms.
Yang for your testimony.
TESTIMONY OF JENNY R. YANG, J.D., SENIOR FELLOW, URBAN
INSTITUTE
Ms. YANG. Thank you. Chair Bonamici, Ranking Member Comer,
and Members of the Subcommittee, thank you for inviting me here
today. I am a Fellow at the Urban Institute, but the views
expressed are my own and shouldn't be attributed to Urban, its
trustees or funders.
I would like to start by sharing a story of Kyle Behm, a
bright college engineering student who applied for an hourly
job at Kroger. He had held similar positions in the past yet
after taking a personality assessment, he was scored red and
rejected.
Kyle had earlier been diagnosed with bipolar disorder, so
personality questions such as whether he experienced mood
changes led many major retailers to reject him.
Sadly, Kyle is no longer with us today, but his father
Roland continues to advocate to ensure people with disabilities
are not systematically excluded by hiring assessments.
A new generation of AI-driven screens are transforming the
lives of America's workers with profound implications for civil
rights. To ensure an equitable future, we must ask the question
who is at risk of being screened out. Otherwise, workers who
fall outside of a set profile could be unemployable for reasons
that are--aren't truly job related. Today, I will focus on two
areas. First, I will discuss algorithmic hiring and
discrimination. Second, I'll address new tech-driven civil
rights concerns for workers on the job. Let's take a look at
the stages of the hiring process through this hiring funnel. In
the sourcing stage, employers recruit applicants. In the
screening phase, employers assess applicant's abilities. In the
interviewing stage, many employers now use video interviews to
evaluate candidates. Finally, employers select candidates and
set pay. In each stage, complex algorithms inform decisions.
Today, I will focus on screening algorithms. Because of the
dramatic rise in online applicants, employees are using chat
bots. Chat bots, resume screens, online assessments and web
games to automate decisions.
Some employers are seeking to increase diversity by
measuring abilities rather than relying on proxies such as
elite university degrees.
Yet many employers simply attempt to automate their past
hiring decisions which may reflect bias. Algorithmic systems
can then replicate existing inequities on a massive scale. Bias
can enter systems in several ways. First, bias may be
introduced in the data using--used to train algorithms.
Amazon's effort to build a resume screen highlights this
challenge.
The computer models trained on resumes submitted over 10
years which were mostly from men. The model then learned to
prefer males and penalize women's resumes containing words such
as women's chess club or all women's colleges.
Second, bias may arise from the variables considered.
Models may learn to utilize proxies for protected
characteristics. For example, zip codes can be a proxy for
race. The selection of variable can reflect the blind spots of
developers, a particularly acute concern given the lack of
diversity in the fields.
Finally, humans may misuse the predictions and place undue
weight on them. To ensure safeguards, I share three strategies
for consideration.
First, an update to the Uniform Guidelines on Employee
Selection Procedures of 1978 would incorporate the latest
scientific understanding into unified government principles.
Second, a third-party auditing system would promote
accountability while having flexibility to evolve with
technology and protect intellectual property.
Third, a worker's bill of rights for algorithmic decisions
would ensure that individuals understand how decisions are made
and have a process to challenge them.
Next, I'd like to turn to new tech-driven civil rights
concerns for workers on the job. One significant concern is
that increased surveillance and tracking of workers'
interactions throughout the day may deter workers from coming
together to raise civil rights concerns for fear of
retaliation.
Another concern is that a growing reliance on customer
ratings by tech platforms and automated performance systems can
introduce harmful and unchecked bias.
Finally, online platforms have disrupted traditional
employment relationships, classifying many workers as
independent contractors.
As non-employees, they aren't protected by most Federal
antidiscrimination laws. Although Section 1981 prohibits
intentional discrimination in contracting based on race and
ethnicity, it doesn't prohibit other forms of discrimination
such as sexual harassment.
States are filling these gaps by providing protections for
independent contractors and making it more difficult to
misclassify workers.
To ensure a future that advances equal opportunity, we need
safeguards that create meaningful accountability. Focus cannot
remain solely on optimizing processes for employers but must
also consider the impact on workers' dignity and civil rights.
Thank you. I look forward to your questions.
[The statement of Ms. Yang follows:]
[GRAPHIC(S) NOT AVAILABLE IN TIFF FORMAT]
Chairwoman BONAMICI. Thank you for your testimony. Dr.
Ajunwa.
TESTIMONY OF IFEOMA AJUNWA, J.D., PH.D., ASSISTANT PROFESSOR OF
EMPLOYMENT AND LABOR LAW, CORNELL UNIVERSITY
Ms. AJUNWA. Chair Bonamici, Ranking Member Comer, and
members of the subcommittee, thank you for the opportunity to
testify today.
I am a labor and employment law professor at Cornell
University and I have been asked to testify today on two
topics.
The first, employment discrimination and privacy concerns
arising from automated hiring including automated video
interviewing. And the second, privacy and discrimination
concerns related to the use of workplace wellness programs and
electronic workplace surveillance.
These technological advancements and the potential for
employment discrimination beg for updates to labor and
employment law.
I identify three major problems with automated hiring. The
first, the design features of automated hiring platforms may
enable employers to eliminate applicants from protected
categories without retaining a record.
Second, intellectual property law which protects automated
hiring from scrutiny could allow discriminatory practices to go
undetected.
And third, the unrestricted portability of applicant data
from automated hiring systems increases the chances of repeated
employment discrimination resulting in algorithmic black
balling.
Automated video interviews are the newest trend in
automated hiring. With this new technology, candidates'
responses are captured on video and then evaluated based on
word choice, speech patterns, and facial expressions.
When video interviewing systems are trained on White male
voices and faces, this disadvantages both racial minorities and
White women whose facial expressions and tone of voice might be
misinterpreted.
Other issues associated with automated hiring include the
unregulated collection of applicants' personal data and the
black box nature of how such information is used.
To date, there are no Federal regulations governing the
collection, storage, or use of data from automated hiring. To
remedy this, I propose three updates to labor and employment
law.
First is the addition of a third cause of action, the
discrimination per say doctrine, to Title VII. Second, the
requirements for audits and certification of automated hiring
systems. And third, a mandate for data retention and record
keeping design features for automated hiring systems.
In addition to automated hiring, technology has advanced
the capability of employers to monitor their workers through
digital surveillance and also employee wellness programs.
Beginning with punch card systems, advancing to GPS systems,
and most recently microchips embedded under the skin, invasive
workplace surveillance is now a part of life for most
Americans.
For example, workplace wellness programs have evolved to
offer health risk assessments and despite protections afforded
by antidiscrimination laws, employers have started to offer
genetic tests to employees.
With the introduction of genetic testing to workplace
wellness programs contradicts both the letter and the spirit of
the Genetic Information Nondiscrimination Act and the Americans
with Disabilities Act.
To protect the health privacy of workers, my coauthors and
I have proposed two new laws. First, the Employee Privacy
Protection Act, the EPPA, would ensure that employee monitoring
is constrained to the workplace and actual job tasks.
The EPPA would limit surveillance outside the workplace and
would prohibit the monitoring of employees when they're off
duty.
Second, the Employee Health Information Privacy Act, the
EHIPA, would clarify that health information generated from
workplace wellness programs are--is protected information under
existing antidiscrimination and health privacy laws.
The EHIPA would also ensure that data collected from
workers could not be sold without the employee's consent.
For the future of work, the primary concern should be
whether workers will enjoy equal opportunity for employment and
also thrive in workplaces that respect human privacy.
Governmental action is necessary to protect workers from
being forced to trade their dignity in the employment
bargaining. I thank the Committee for the opportunity to
testify today and I look forward to your questions.
[The statement of Ms. Ajunwa follows:]
[GRAPHIC(S) NOT AVAILABLE IN TIFF FORMAT]
Chairwoman BONAMICI. Thank you for your testimony. I now
recognize Ms. Lander for 5 minutes for your testimony.
TESTIMONY OF ESTHER G. LANDER, J.D., PARTNER, AKIN GUMP STRAUSS
HAUER & FELD LLP
Ms. LANDER. Thank you, Chair Bonamici, Ranking Member
Comer, and Members of the Subcommittee for allowing me to
appear before you today.
I am a partner at the law firm Akin Gump in the firms'
labor and employment group here in Washington, D.C. I
previously served as the Principal Deputy Chief in the
employment litigation section of the Department of Justice
Civil Rights Division and I am appearing here today in my
personal capacity.
In my written testimony, I describe the many benefits
associated with using technology in employee selection
procedures.
If used correctly, the business case is clear. Employers
are able to harness the power of available data to efficiently
make sound selection decisions, reduce manual labor, subject
candidates to the same objective screening criteria, and
eliminate the potential for implicit bias that exists with
subjective decision making.
With so many technology-based tools on the market however,
concerns have been raised that AI screening is resulting in
unlawful discrimination.
To date there have been few lawsuits challenging AI tools
and there are no published studies to show technology-based
selections are more likely to result in discrimination than
more traditional paper and pencil tests.
With that said, when employers implement technology to make
selection decisions, it is important to understand the laws
that already exist to protect applicants and candidates from
unlawful discrimination.
Specifically, Congress passed the Civil Rights Act of 1991
which amended Title VII to make disparate impact discrimination
an unlawful employment practice.
Under the 1991 act, any selection procedure that adversely
impacts protected groups must be justified by the employer as
job-related and consistent with business necessity.
To make this showing, employers must document a strong
connection between the selection procedure and the job in
question which typically involves a process called testing
validation.
Courts assess the adequacy of an employer's validation
efforts under the Uniform Guidelines on Employee Selection
Procedures which were adopted by the EEOC and other government
agencies to assess the lawfulness of selection procedures under
Title VII.
Although the guidelines were established in 1978, which
admittedly was a long time ago, they are well equipped to
address the concerns expressed by other witnesses today about
AI tools resulting in hiring decisions based on non-job-related
correlations that screen out protected groups.
First, the guidelines anticipate developments in hiring
techniques and tools and make clear that all selection
procedures need to be reviewed in light of current
understandings which in itself is a basis to reject validation
studies premised on non-job-related correlations.
Second, the guidelines direct enforcement agencies to
consider whether the selection procedure was carefully
developed and is being used in accordance with professional
standards. This concept is commonly referred to as competent
test design.
So for example if an AI tool has machine learned to
disproportionately screen out applicants from a protected group
because they do not share the same zip code as successful
incumbents, an employer would not be able to show competent
test design even if a strong correlation exists between
performing successfully on the job and one zip code. Third, the
guidelines require that all validation studies include a
complete and explicit description of the selection procedure
that includes any measures that are being used. This written
transparency requirement means that vendors cannot hide behind
the so-called black box.
A proper validation study that complies with the guidelines
must explain what the selection procedure is measuring and then
correlate those measures with successful job performance,
reduced turnover, or other important job-related behaviors.
And finally, regardless of how a selection procedure is
validated, the guidelines require an investigation into
fairness.
This investigation could include taking a deeper look at
the selection procedure to see what items were in the case of
AI tools, which screening criteria are causing adverse impact
and to consider removing those criteria and making other
modifications that will result in a fair selection procedure.
I'd also like to briefly address the gig economy, an area where
advances in technology have created opportunities for works and
companies.
Gig workers can take advantage of low costs, flexible
hours, and the ability to easily build an independent business.
The ease of technology and the volume of workers using it has
heightened concerns about worker misclassification.
However, there is a body of law that already exists to
address this topic as does a comprehensive remedial scheme for
workers who have been misclassified. The remedies for
misclassified workers grow even more substantial when recovered
on a class-wide basis which have served as a powerful deterrent
against worker misclassification.
In closing, technology advances are beneficial to workers,
employers, companies, and the economy. As the labor force and
businesses adapt to these changes, employment laws are
currently in place to ensure that worker rights are protected.
Thank you for the opportunity to speak with you today and
share my thoughts on the important topics covered by this
hearing. I look forward to answering your questions.
[The statement of Ms. Lander follows:]
[GRAPHIC(S) NOT AVAILABLE IN TIFF FORMAT]
Chairwoman BONAMICI. Thank you for your testimony. I now
recognize Mr. Romer-Freidman for 5 minutes for your testimony.
TESTIMONY OF PETER ROMER-FRIEDMAN, J.D., PRINCIPAL AND HEAD OF
THE CIVIL RIGHTS AND CLASS ACTIONS PRACTICE, GUPTA WESSLER PLLC
Mr. ROMER-FRIEDMAN. Thank you. Good afternoon and thank you
for the opportunity to testify today. My name is Peter Romer-
Friedman, I'm a principal at Gupta Wessler PLLC and the head of
the firm's civil rights and class actions practice.
As a civil rights lawyer, I've represented victims of
discrimination in jobs, housing, credit, and public
accommodations. They've included workers in many industries,
service members, and veterans, victims of Hurricane Katrina and
the foreclosure crisis, as well as farmers and ranchers. Lately
I have focused on combatting digital bias.
Sixty years ago, there were no desktop computers or
websites, but we did have entrenched discrimination in the
workplace, housing, in public spaces.
If you picked up a newspaper in 1960, you'd see classified
ads with segregated columns for male and female jobs. Job ads
that stated explicit preferences based on race, gender, and
age.
Congress tried to put an end to this biased advertising and
recruiting when it enacted Title XII of the Civil Rights Act
and the Age Discrimination in Employment Act.
Congress knew this discrimination has huge negative
consequences. If you announce a job is for men, women are less
likely to apply. If you primarily recruit men, mostly men will
be hired.
For decades it appeared that these laws were working. Overt
discrimination and statements in newspapers disappeared. Most
employers stopped openly recruiting based on biased
preferences. This all changed however, when employers decided
to harness the power of the internet and social media to
recruit workers. Advertising platforms like Facebook enabled
employers to discriminate in their job advertising so that they
could target job ads only to people of certain races, genders,
ages, zip codes, and even political interests.
An untold number of employers deployed these very tools to
expressly exclude workers from receiving their job ads based on
many protected traits.
And until recently when Facebook made changes due to a
settlement with my clients, it was possible for employers to
exclude people from getting job ads based on thousands of
categories unrelated to jobs.
For example, an employer could decide not to send their job
ads on Facebook to people interested in Christianity, the
Republican National Committee, the ACLU, or the AFL-CIO.
And just a few years ago, employers could target job ads on
Facebook to people interested in heinous things like Hitler,
White pride, fascism, rape, and ISIS.
There has never been a full public accounting of all the
biased ads published on Facebook but here is what we know from
investigative journalism and the investigation of my client,
the Communications Workers of America.
Hundreds if not thousands of employers routinely excluded
women and older workers from getting job ads on Facebook. The
same bias was common in ads for housing, credit, and other
financial services.
There have likely been hundreds of millions of incidents of
digital bias. Here are a few real-life examples. T-Mobile sent
job ads on Facebook targeting people who were only 18 to 38
years old. Amazon sent job ads on Facebook that targeted only
people 18 to 50. A leading security installation company called
Defenders sent job ads targeting only men 20 to 40.
Thankfully, many terrific advocates stepped up to challenge
this harmful discrimination. Organizations like the CWA, the
ACLU, National Fair Housing Alliance and my prior law firm,
Outten & Golden.
We took Facebook to court and filed EEOC charges against
dozens of employers that denied job ads to women or older
workers.
After years of litigation, Facebook in March of 2019,
agreed to make sweeping changes to its platform to prevent
advertisers from denying job, housing, and credit ads based on
protected statuses and Facebook recently implemented those
changes. Still, we are very concerned that Facebook's own
algorithm may be discriminating based on age and gender when
Facebook itself decides which users will receive job ads within
an audience the advertiser selected.
We are also troubled that dozens of major employers
including Amazon, T-Mobile, and Capital One are claiming that
Federal law does not bar them from denying job ads to workers
based on a protected status like age.
We believe our Federal civil rights laws already outlaw
this crude digital bias and recently we have seen the DOJ,
EEOC, and HUD agree that it's illegal to deny job or housing
ads based on a person's race, gender, or age.
But Congress can and should take critical steps to clarify
and strengthen Federal law to stop digital bias. I have
recommended a range of critical steps that Congress can take
including ensuring that tech platforms like Facebook are
covered by civil rights laws, clarifying that certain types of
digital bias are unlawful, requiring greater disclosure of
digital practices and bias, and making sure that the Federal
public accommodations law applies to online spaces, and ending
section 230(c) immunity for commercial or paid advertising.
In too many areas of our society, the move fast and break
things credo of powerful technology leaders like Mark
Zuckerberg has turned back the clock by more than half a
century. It has upended our civil rights, our civil discourse,
and even the most basic facts that our society can agree upon.
Technology should not disrupt our civil rights. It
shouldn't break equal opportunity. Technology should be a
mechanism for making the promise of equal opportunity and
integration a reality, especially in the workplace.
Thank you very much, appreciate the opportunity to answer
any questions.
[The statement of Mr. Romer-Friedman follows:]
[GRAPHIC(S) NOT AVAILABLE IN TIFF FORMAT]
Chairwoman BONAMICI. Thank you so much to each of our
witnesses today. Under the Committee Rule 8(a), we will now
question witnesses under the 5 minute rule.
And I want to say in light of all the testimony we heard
today, I'm sure everyone wishes for more than 5 minutes because
we have so many questions but I will now yield myself 5
minutes.
Professor Ajunwa, in your written testimony, you discuss
how companies use automated video interviewing that permits the
employer to evaluate factors that are not job related in the
interviewing process.
Last year, in the Science, Space and Technology Committee,
Joy Buolamwini is the founder of the Algorithmic Justice League
testified on some of these issues and they discussed their
experience with facial analysis software failing to detect
their dark skin until they put on a white mask which uncovered
both skin type and gender bias in the AI services from
companies like Microsoft, IBM, Amazon.
So, Professor, what characteristic can employers evaluate
when using automated video interviewing and do individuals
typically know these factors being evaluated as they interview?
Ms. AJUNWA. Thank you for your question, Chair Bonamici.
So, the--one of bigger problems that automated video
interviewing is that oftentimes the job applicants don't
actually know that they will be evaluated based on their video.
They just think that they're sending in a video that will
then be viewed by humans but actually that video is actually
being put through algorithms that are evaluating both the
facial expressions, tone of voice, even word choice.
And the problem with that of course is that if you look at
how the training of the algorithm is done, oftentimes the
training is using a very limited pool of applicants so it could
be all White male applicants and in which case, women who have
different tones of voices or even people who are from other
cultures and therefore have different facial expressions can
actually be disadvantaged because then their responses can be
misinterpreted by the algorithm.
Chairwoman BONAMICI. Thank you. Both on this Committee and
on the Science Committee, we have a lot of conversations about
the importance of diversifying the STEM, STEAM workforce and I
think that is one step in solving this problem because
obviously, it is designing the algorithm it makes a difference.
Mr. Romer-Friedman, in your testimony you said that job
advertisements are often targeted based on categories that are
not job related, or proxies, and you described how individuals
may be excluded from seeing job ads. Thank you for the actual
visual representation.
Do you consider excluding an individual from seeing a job
ad because their experience exceeds a maximum number of years
or because they attended a women's college for example, would
those be examples of targeting based on proxies?
Mr. ROMER-FRIEDMAN. Absolutely, Chair Bonamici. These are
the kinds of things that without the digital procedures and
processes could be illegal. We see this a lot in the economy
for older workers. They're excluded simply because they have
too many years of experience or they graduated from college a
number of years ago.
But we are seeing the something accelerated and exacerbated
in the digital space and that's a problem. We think that it
clearly violates the law. It not only has a disparate impact;
we think you can infer intentional discrimination from these
kinds of clear proxies.
Chairwoman BONAMICI. And how could Congress best make sure
that the employers are not using proxies to discriminate based
on sex, age, religion, other categories, protected categories?
Mr. ROMER-FRIEDMAN. As I have recommended in my testimony,
Congress could explicitly say that if a job category or if a
category for targeting someone or evaluating someone is not
directly related to the job or the opportunity, it simply is
banned, it would be an unlawful practice.
In the same way that it's just strictly unlawful regardless
of the intent to advertise a job that states a preference based
on age or race or gender.
Chairwoman BONAMICI. Thank you. Ms. Yang, you're a former
Chair of the Equal Employment Opportunity Commission. What
additional resources could Congress provide to the EEOC, the
Commission on Civil Rights, and the Office of Federal Contract
Compliance Programs to adequately address the problems that
were described today with algorithmic bias and digital
discrimination in hiring?
Ms. YANG. Thank you, Chair Bonamici, for that question. The
government plays a particularly important role in rooting out
hiring discrimination because individuals typically don't know
why they weren't hired.
So the EEOC made it a priority to look at recruiting and
hiring discrimination and the agency has authority to open
charges on their own investigation, even where an individual
may not have enough information.
So right now under our current law, the Federal Government
plays an incredibly important role in investigating concerns
about hiring screens and the agencies need more resources. They
need to be able to hire computer scientists and data scientists
who understand how these systems work.
We had initially started a task force over 4 years ago back
when I was at the EEOC. We had Professor Ajunwa testify and
help us learn about these issues but we didn't have the
capacity on staff to really, fully understand how to evaluate
these systems, to understand how the Uniform Guidelines really
need to be updated and having that technical know-how within
the agency would be incredibly valuable.
Chairwoman BONAMICI. Thank you so much. I yield back and
recognize the Ranking Member of the Full Committee, Ms. Foxx,
from North Carolina for your questions.
Ms. FOXX. Thank you, Madam Chairwoman, and I want to thank
our witnesses for being here today. Ms. Lander, the Federal
laws prohibiting employment discrimination do not explicitly
address the technologies we are discussing today. In your
opinion though, are these statutes readily applied in the
modern workplace and to employers' use of search engines,
algorithms, and AI in the recruitment and screening process? Do
you see gaps in these laws or do you believe these laws are
more than broad enough to cover new technologies?
Ms. LANDER. Thank you for that question. As a practitioner
in this area who actually has counseled clients and reviewed
some of the tools that we're talking about today, I have not
had any difficulty applying the Uniform Guidelines as written
to assess these tools and to provide feedback to both my
clients and the vendors who are selling them regarding ways in
which they should be modified or enhanced to ensure
nondiscriminatory selections.
So to answer your question, yes, I do believe that it is
not difficult to apply the Uniform Guidelines as they're
currently written to address the concerns that are being raised
by the panel today with regard to the technology tools that are
on the market.
Ms. FOXX. Thank you, Ms. Lander. Ms. Lander, what are the
upsides for business owners of using new technologies in
recruiting and screening job candidates from the perspective of
complying with non-discrimination laws?
In what ways are these technologies superior to other forms
of job recruitment and screening when it comes to complying
with the Federal laws prohibiting discrimination?
Ms. LANDER. Well, in this current climate that we are in,
the volume of resumes and applications that are submitted, it's
quite different than the day and age where somebody had to walk
in a fill out a paper application.
Employers are being bombarded with, you know, thousands of
applications sometimes when they have an opening and in its
simplest form, AI tools are capable of simply scanning those
applications or resumes simply to screen out those who don't
even have the minimum qualifications for the job which saves a
substantial amount of man power and time in trying to do that
with a person.
To answer your question though about how these tools can
help reduce discrimination, when done correctly, these tools
eliminate the risk of implicit bias in decision making because
when the criteria that they're screening for are job related,
the entire screen is objective and is not susceptible to what
somebody might believe when they see a particular name or they
look at somebody and see a particular race or gender.
Ms. FOXX. Thank you. Ms. Lander, you discussed in your
written testimony, the Uniform Guidelines on Employee Selection
Procedures jointly written by the EEOC, the Civil Service
Commission, Department of Labor, and Department of Justice and
provide guidance for the employers and obviously you have
mentioned those in your comments now.
Based on your experience, do the guidelines apply to the
algorithms in AI that many employers are using that is somewhat
repetitious to my first question, and do the guidelines provide
useful information and best practices for employers?
Ms. LANDER. They do. Some of the tools that we're
discussing are recruiting tools and some are hiring or
selection tools.
So the guidelines are aimed at any sort of hiring test or
selection device that makes decisions that allows people to
proceed in the hiring process.
So when it comes to recruiting, there is a difference
between sourcing, which is efforts to expand your pool of
eligible candidates or applicants that meet your qualification,
the qualifications for the job.
And so some of these tools are being used to simply to
expand and enhance traditional forms of recruiting. And if
people aren't being excluded, if it's not an exclusive method
of recruiting or the sole method of recruiting, then arguably
adverse impact is not going to be an issue and so in that case,
if there is no adverse impact, the guidelines don't come into
play.
Ms. FOXX. Quick, quick question. If an employer wrongly
classifies an employee as an independent contractor, isn't
there significant potential liability for the employer
including back pay and liquidated damages under the Fair Labor
Standards Act which provides substantial incentives not to
classify workers incorrectly?
Ms. LANDER. Yes. That's correct. Not only is back pay and
liquidated damages available, for willful violations, the
statute of limitations goes back 3 years.
And when you--the class action activity in this space has
been quite active and it has actually changed behavior and a
lot of employers and companies and workers are all quite aware
of the issues involved with misclassification and rights are
being protected and asserted on a regular basis through the
courts.
Ms. FOXX. Thank you, Madam Chairwoman, I appreciate your
indulgence.
Chairwoman BONAMICI. I now recognize the Chairman of the
Full Committee, Mr. Scott from Virginia for 5 minutes for your
questions.
Mr. SCOTT. Thank you. Ms. Ajunwa, several of you mentioned,
you know, if a women's college gets mentioned that could have a
negative effect. Who decides what instructions are given to
create the algorithm and what happens when you get some kind of
hit? Who does it, who actually designs it?
Ms. AJUNWA. Thank you very much for your question. I guess
the question is who comes up with the criteria used for
programming the algorithm.
And oftentimes algorithms are programmed by vendors who
then sell them to employers. But also, employers can have
algorithms that I call bespoke, meaning that these algorithms
are created specifically for that employer.
So this of course can change the liability whether the
algorithm has been created by the vendor or specifically
created at the behest of the employer.
So I guess my focus today is really the first scenario.
When the algorithm is being created by a vendor and the
employer perhaps does not know exactly what has gone into the
algorithm and also how it has been trained.
So really, my advocacy today is really for both the
auditing and certification of automated hiring systems before
they are deployed, before they can actually be used in the
workplace. Because I do believe that, you know, as all the
witnesses have stated, if automated hiring is used properly or
correctly, they could be helpful.
The problem is they currently are not, right. The problem
is that there are currently no regulations to actually ensure
that they are being used correctly and appropriately.
Mr. SCOTT. Well, once you have designed--once its designed
with the discrimination kind of embedded, if the employer
bought it from a vendor, would they be immunized from any kind
of intentional discrimination?
Ms. AJUNWA. So that is a gray area. That's a gray area in
terms of the law because one thing that Title VII requires is
intent and other than intent the showing of disparate impact.
And both things can be hard to prove if the automated
hiring system is coming from a vendor because first, you can
argue perhaps there is no intent on the side of the employer
but then there is also the issue of even establishing disparate
impact because you would need statistical proof and the
automated hiring is--the automated hiring system might have
been designed not to retain all the record that you need for
that group.
Mr. SCOTT. Well, how does the employer know, he buys this
little algorithm thing and uses it, turns out it's screening
people. How would he know?
Ms. AJUNWA. He wouldn't. So that is why I am advocating for
an audit requirement for employers who do then buy automated
hiring systems or use automated hiring systems.
Mr. SCOTT. Thank you. Ms. Yang, the Ranking Member brought
up independent contractors. If you are an independent
contractor, you are not protected under the employer employee
Title VII, ADA, and others.
In--but you would be protected under Section 1981 of the
Civil Rights Act of 1866 where you can't discriminate. Are
there limitations in Section 1981 in terms of pursuing
discrimination claims if you are an independent contractor?
Ms. YANG. Yes. Federal law provides very limited
protections for independent contractors. Under Section 1981,
the claims must prove intentional discrimination which can be
very difficult to show in the case of algorithmic bias.
In addition, it only covers race and also ethnicity
discrimination but not other bases, right. So it wouldn't cover
sexual harassment, age discrimination, disability--based
discrimination.
And our Federal antidiscrimination laws contemplate that
true independent contractors will have the bargaining power
that they don't need to be protected against discrimination.
But the way in which many companies are misclassifying
independent contractors today means that many individuals do
not truly have bargaining power and they need the protection of
our antidiscrimination laws.
Mr. SCOTT. Thank you. Madam Chair, I yield back.
Chairwoman BONAMICI. Thank you, Mr. Scott. I now recognize
Ms. Stefanik from New York for 5 minutes for your questions.
Ms. STEFANIK. Thank you, Chairwoman Bonamici. Ms. Lander, I
appreciate that you raised how contractual or gig arrangements
can be beneficial to workers, as I believe this perspective
needs to be central in our discussions on worker
classification.
Millennials, as you know, now comprise the largest cohort
in the U.S. labor force and these workers place a higher value
on the flexibility and fulfillment that can exist outside the
rigid constraints of traditional employment.
For years, independent contracting has sparked
entrepreneurship and provided an important source of income and
flexibility to millions of Americans including students,
veterans, and single parents.
In your testimony you mentioned how there are various legal
tests courts and government agencies applied to distinguish
employees from independent contractors.
I have heard from employers, particularly small business
owners, that this inconsistency between various agencies has
muddied the line on worker classification and really created
compliance challenges.
Do you believe that harmonizing the legal test across
Federal agencies would help draw a clearer bright line on the
issue of worker classification and help workers as well as
business owners know when misclassification has indeed
occurred?
Ms. LANDER. Yes, I do think that would actually make life a
lot easier for employers. However, the problem is that the
definition of employee differs from statute to statute and so
unfortunately what that means is when courts are interpreting
whether a particular civil rights or labor law applies, they
have to look at the statutory text and apply it. So as easy as
it would be to have a uniform definition, if you're going to be
honest to the statutes that involve workers, you can't have a
uniform definition across all of the agencies.
Ms. STEFANIK. So how would you address that then? If there
is a uniform definition legislation which I have worked on,
what would we need to do in addition to that?
Ms. LANDER. Well, I'm not a lawmaker so I can't really
answer that question.
Ms. STEFANIK. Great. Well, your perspective is important on
that. I would like to follow up on that issue to make sure that
we get this right.
And very briefly, what would happen, what would be the
impact of bringing California's ABC test nationwide and would
it allow workers who value freedom and flexibility the choice
to maintain their status as independent contractors?
Ms. LANDER. My understanding of the California test is that
it moves away from the traditional right to control which is a
critical element in all of the independent contractor analyses
under the various laws and talks about the essence of the
business.
And so if a worker is engaging in services that is the
essence of the company's business and I don't know if I'm
wording that exactly right, then he or she can't be an
independent contractor.
And that would essentially completely change the entire
working dynamic for not just the gig economy which has been a
tremendous boon and not only for companies that have been able
to expand their reach where they otherwise couldn't, but it's
also been wonderful for as you described in your opening
remarks, for individuals who need the flexibility to work
different schedules and seasonally and things of that sort.
Ms. STEFANIK. Thank you. You know, as we discuss this
issue, I think it is really important that we channel these
technological and entrepreneurial opportunities for young
people and members of the nontraditional workforce, people that
maybe are augmenting their full time job, people that as they
are aging want to earn some money on the side. You know, there
is lots of benefits to this gig economy and we have to
remember, it is totally voluntary by individuals who seek out
those opportunities. And with that I yield back.
Chairwoman BONAMICI. Thank you very much and I now
recognize Dr. Schrier for 5 minutes for your questions.
Ms. SCHRIER. Thank you so much to all of you for being
here. I really enjoyed reading your testimonies and hearing you
today and it is so interesting to think about really there is a
rabbit hole that you can go down when you start thinking about
how every question you ask or every parameter you put in an
algorithm can lead down the line to some sort of
discrimination.
And I think that this was all developed for efficiency and
to cast like a broader net but a more specific net but in doing
that with so many of the things you talked about like age or
even look alike that looks at a current workforce, has led to
discrimination inadvertently.
And so I represent part of Washington State and they just
had a future of work task force and they released a report in
December talking about automation in the workplace and how AI
will change the way we work but it barely touched on this topic
of algorithmic discrimination and how that leads to people even
finding out about jobs or being eligible for jobs. And so, Ms.
Lander, you talked about kind of a look back, you know, once a
system is in place, how do we look at it and see if it is
discriminating.
And I am wondering if there is a way to look forward? So
this is sort of question for Ms. Yang, Dr. Ajunwa about whether
you--whether there are things that we can do to either
fundamentally change the way these algorithms work or whether
we should look in another place and change privacy laws so that
the algorithms can't even obtain some of that information and
how you might balance those two.
Ms. YANG. Thank you for that question. I think we have a
lot of opportunity to make algorithms work more fairly than
they are right now. And it starts with ensuring that the
information you're considering is truly job related.
And we talked about the training data. Is it diverse and
representative of the full spectrum of people that can perform
the job? And then what are the criteria that you're building
into the variables? Are you thinking about abstract personality
characteristics that maybe have some correlation but a heck of
a lot of people would also be able to perform the job even if
that weren't their top personality characteristic??
And so it comes back to ensuring that we are really being
rigorous about the screen being job related. And the closer you
can tailor what you're selecting for to behaviors on the job,
the more you can minimize the risk of screening people out who
could perform the job.
And I do think many advances in technology now will allow
us if you design a system up front to document the decisions so
that you can explain how they were made which is necessary
under our current laws to ensure accountability. Employers
themselves, even if they say I didn't design it, I didn't know
what was in the algorithm, they are nonetheless responsible.
And my view is that they absolutely need to understand how
these decisions are made. They need to be able to explain them
through when the government comes in and asks about their
system or in litigation and I do think we need new laws both to
protect privacy but also to create the right incentives because
these cases are very expensive to litigate.
Ms. SCHRIER. This is, I'll get back to kind of a follow-up
question. I wanted to give you a chance Dr. Ajunwa to give a,
your answer and then I have a follow-up question about it.
Ms. AJUNWA. Thank you very much for your question. I do
strongly agree that we have to be forward looking because being
backward looking is basically taking action after the harm has
already been done and I think we can actually prevent a lot of
harm from the onset.
And that includes for example mandates for the design of
these automated hiring systems which we don't yet have. And
you're very right to pinpoint that part of the problem is the
way that we handle privacy, especially privacy of workers in
the United States and that part of the problem is thinking
through what sort of information is actually being pulled into
the system of automated hiring--
Ms. SCHRIER. And it is all out there.
Ms. AJUNWA. Right.
Ms. SCHRIER. Can I just quickly in the interest of time, my
next part was about you had said are these issues kind of
pertinent to the job?
And so a few years ago, Google had a project called
Aristotle and they found out that what really mattered was not
so much your engineering degree but how well you worked with
others.
Ms. AJUNWA. Right.
Ms. SCHRIER. And so they kind of lifted up characteristics
like a team leader or a club leader or being on a sports team.
But even that then chooses for perhaps competitive people or
people who always want to be the star of the show and might not
really lead to the best workplace.
I wondered if you could just comment about that because it
is job related but it could have inadvertent outcomes.
Ms. YANG. Part of the challenge is that you may be testing
only on your current workforce, right. So you will be
replicating that current model.
I think algorithmic formulas, you know, algorithmic systems
can help us identify bias within broader systems. You know, we
might think confidence, you know, expressed in resumes as words
like executed, will mean you're going to perform well. In fact,
more men use those words and in fact that might not mean you
can perform well, right. Confidence doesn't always equal
competence.
And I think the more we can use these kinds of technology
systems to help identify where some of the bias is within
processes, then we can actually start to break down some of the
historic bias.
Ms. SCHRIER. Thank you.
Ms. AJUNWA. And I would add that, you know, having actual
record-keeping mandates would aid in this endeavor, right. So
to be able to see what are the people that are actually
applying, what are the people that are getting selected, but
then also checking that against the wider pool that's out
there.
So, you know, somebody mentioned nontraditional workforce.
So for example, people who have gaps in employment are
oftentimes excluded algorithmically by automated hiring
systems. And this can negatively impact women who are often
called upon to be caregivers.
It can also impact formerly incarcerated citizens who have
been rehabilitated and who are trying to reenter the workforce.
It can impact veterans.
So I really think, you know, having a proactive approach to
ensure that there is proper record-keeping for automated hiring
systems and also proper auditing of automated hiring systems
will really be a boon for employers, not just employees.
Ms. SCHRIER. Thank you.
Chairwoman BONAMICI. We are going to move on to the Ranking
Member of the Subcommittee, Mr. Comer from Kentucky, for his
questions.
Mr. COMER. Thank you, Madam Chair, and I appreciate all the
witnesses being here today. Ms. Lander, in the modern economy,
job recruitment is migrating online.
Based on your experience, what are job seekers and
employers gain from the use of online platforms when it comes
to finding and filling jobs?
Ms. LANDER. The ability to scan the internet to find
opportunities for work is a tremendous gain for workers. I can
remember back when I was job hunting and had to look in the
newspaper at classified ads so it's a completely different
world that we live into today.
Mr. COMER. Ms. Lander, under current law if an employer is
using technology to screen job applicants that has a negative
impact on a protected class, the employer may need to
demonstrate the screening criteria is job related and
consistent with the business necessity.
What goes into conducting this analysis and would the
employer have to demonstrate a strong connection between the
screening criteria and the job that the employers trying to
fill?
Ms. LANDER. Yes. It's the Uniform Guidelines process for
validating a selection device is extremely rigorous. The two
most common ways are content validity and criterion validity.
Content validity is less likely to apply to the kinds of
tools that we're discussing because content validity is
typically the content of the test or selection device matches
the content of the job, like a pilot simulator or a typing
test. Here, we are talking about devices that screen for either
minimum qualifications or particular personality
characteristics and those are typically justified by criterion
validity which is a rigorous statistical process of matching
performance data with performance on the selection device.
Mr. COMER. So what are some best practices for an employer
when it is considering using an online platform or a vendor
that employs AI to find suitable job candidates?
Ms. LANDER. As Ms. Yang said, the employer can't get off
the hook simply by saying that the employer relied on the
vendor.
So employers are responsible for knowing how they are
screening their candidates and so any employer that is thinking
of using a tool that uses AI or any other sort of technology to
screen candidates should be insisting upon seeing the vendor's
adverse impact studies as well as the validation work that has
been done and to understand what kind of screening criteria is
being used to screen their candidates.
Mr. COMER. Okay. Let me ask you this one last question. If
an employee is incorrectly classified as an independent
contractor, wouldn't this worker retain all the legal
protections of an employee including the protections of our
current civil rights law?
Ms. LANDER. Yes. Misclassified workers who are actually
employees are protected by all of the employment laws.
Mr. COMER. And I want to ask Mr. Romer-Friedman, you had
mentioned Facebook in your opening testimony and what--Facebook
gets a lot of criticism, bipartisan criticism, here in
Congress. What can Facebook and what should Facebook do
differently with respect to this subject we are talking about
here today?
Mr. ROMER-FRIEDMAN. Sure, thank you, Ranking Member Comer.
So we have already made a lot of progress lately with Facebook.
They've created a special portal for job, housing, and credit
ads where you don't have at this point in time most of these
demographics as selection options to target or exclude and
that's great and we applaud them for doing that. At the same
time, as I mentioned in my testimony, Facebook has to decide
who will see what ad, right. So let's say I want to send an ad
to everyone here in District of Columbia, but I'll only buy
10,000 impressions, right. 10,000 people who are going to see
the ad.
Facebook has got to decide who is going to see those ads.
We allege and we are going to get this hopefully in discovery
and litigation that age and gender are being used and a group
called Upturn has done a study showing that there are racial
and gender impacts so that even if the employer doesn't want to
discriminate by relying on the ad delivery algorithm of
Facebook, it may be doing just that and even worse than what
was going on for years in the first place where the employer
was expressly excluding certain groups.
And so as you said, and I completely agree with this. Most
businesses want to follow the law and they want to comply.
Mr. COMER. Right.
Mr. ROMER-FRIEDMAN. And that's where I disagree with Ms.
Lander that creating greater clarity in the law always helps
compliance and reduces litigation. And I think that's, you
know, everyone can agree that those are good things.
Oftentimes you do that in regulations that the EEOC can
issue but Congress can do that too and I think the laws that
Congress enacts express the values of this Congress.
So for example, one very basic thing is Amazon says it has
a right to send job ads to younger people and not send them to
elderly people as long as they put the job ad on their website.
That's something where there, they say there's an ambiguity
in the law. Congress could step right in there and make it
clear you can't use race, gender, age, disability, veteran
status, political status, for example to exclude people from
getting recruited or getting job ads. Simply put.
Mr. COMER. Madam Chair, I have to throw this statistic in
here. My congressional district, the recent poll, they polled
all the congressional districts on Facebook usage. 84 percent.
Mine was the second highest in Congress. 84 percent of the
adults in my congressional district get on Facebook at least
once a day. So I am a, I represent a Facebook district so.
Chairwoman BONAMICI. That is fascinating, Mr. Comer.
Mr. COMER. We are also trying to--
Chairwoman BONAMICI. You made me want to--
Mr. COMER. That is right.
Chairwoman BONAMICI.--look at where mine is and everybody
else's is. It is really interesting. And next we recognize Ms.
Lee from Nevada for 5 minutes for your questions.
Ms. LEE. Thank you. Thank you all for being here. This has
been really an interesting topic to think about all the
iterations of what can be, what we view as helping us in this
modern day to actually promoting discrimination that we have
not thought about.
I am going to turn to older Americans because this body
passed the Protecting Older Americans Against Discrimination
Act last month with bipartisan support which restores the
ability of older Americans to apply the so-called mixed motive
framework which was afforded to protect other classes of
individuals under Title VII of the Civil Rights Act to claims
of age discrimination.
So in light of the new challenges that we are facing in the
digital age, I would like to ask you, Mr. Romer-Friedman, you
touched upon this a little bit in your last answer.
Are there other actions that we should be taking to ensure
that older workers have the same protections as other protected
classes?
Mr. ROMER-FRIEDMAN. That's a great question and, you know,
I think my former colleague David Lopez who is the general
counsel of the EEOC during the Obama Administration has pointed
out to me many times that age discrimination has become so
normative and so kind of baked into our society that people
don't even think it's illegal. Right. So we have to I think
treat it very seriously.
To that end, mixed motive is so important to protect
because it this algorithmic bias, digital discrimination
discussion, companies will say well, age was just one of
hundreds of factors that could have influenced that decision.
Of course, then it's very difficult to piece together how
age was used. You shouldn't have to show that age was more
determinative in a decision than you would have to for gender
or race but that's the case right now.
I think, you know, one thing that could be done is making
clear that the Age Discrimination Employment Act applies to
applicants for disparate impact claims. Right.
Two courts of appeals have held that if you want to bring a
disparate impact claim, you can only do it as an employee under
the ADEA, you can't do it as an applicant. And that's the whole
purpose of the ADEA, to allow older people to get hired. And,
you know, at the end of the day we need to make sure that
things like companies not being able to screen out when they're
recruiting based on the date of graduation or the years of
experience and just completely take that person out of the
picture digitally, those are the kinds of things that need to
be implemented right away.
Ms. LEE. Thank you. Thank you. As we talked about this and
I, we deal with this a lot in Congress is that new technologies
are far outpacing our ability to focus on regulating and
certainly that is what we are seeing here today.
So as we look at, I would like to just open this up to all
of you. Looking down the road, are there potential future
developments in workplace or hiring technologies beyond the
ones we have talked about today that particularly concern you
when it comes to protecting workers rights from employment
discrimination? I will start with you, Ms. Yang.
Ms. YANG. Thank you for that question. I am concerned about
the increasing worker surveillance and monitoring. Many workers
now are tracked all throughout the day. There are productivity
metrics that can sometimes be so aggressive that they can
interfere with a pregnant woman's ability to go to the
bathroom, you know, prayer time, all kinds of civil rights
concerns.
But also just the simple tracking of people throughout the
day may really deter workers from coming together and raising
concerns so I do have concerns about that.
And I did want to add one other point about the age
discrimination. You know, older workers are disproportionately
represented at independent contractor positions and so it's
especially important that even properly classified independent
contractors have antidiscrimination protections. Right.
And if somebody says well, I'm not going to hire you just
because you're old, like right now you have no protections
against that. And I think that's something that needs to change
as well.
Ms. LEE. Right. Thank you.
Ms. AJUNWA. Thank you very much for your question. First in
response to, you know, your concern for older workers, I do
want to note that I have seen in my research more
discriminations against older workers in terms of their ability
to participate in a sort of a digital workplace. So people will
use words like digital native to really exclude older workers
so that's something of concern.
I also wanted to point out that workplace surveillance is
actually something that is on the rise. As I mentioned, the
microchips that are being embedded under the skin, but also, I
see workplace wellness programs as a site of workplace
surveillance.
For example, now with the sort of trend or introduction of
genetic testing as part of workplace wellness programs. That
really raises the question of increased, you know, health
discrimination or increased discrimination against people with
disabilities whether real or imagined.
Because genetic testing is actually just telling you the
propensity for disease, but employers might look at it as
actually determinative when it's really not.
So I think that's a huge concern and something we should
really, you know, act against.
Ms. LEE. Right, thank you. All right. My, whoops, my time
has expired. Thank you.
Chairwoman BONAMICI. We now recognize Mr. Takano, a Member
of the Full Committee from California for 5 minutes for your
questions.
Mr. TAKANO. Thank you, Chairwoman Bonamici, for this very
important hearing. As the workforce is changing and we
transition to a society more dependent on technology, it is
extremely important that we understand how these tools will
impact the workforce.
Currently there is a lack of transparency and without
knowing the algorithm behind the program we have no way of
knowing if these tools will remove or reinforce bias.
Professor Ajunwa, as companies are looking to ensure that
they remove bias and are mitigating against disparate impact,
they would need to know what protected classes potential
employees belong too.
We know that the more sensitive the information is,
information such as sexual orientation or disability status,
the less likely a candidate will disclose this information. So
my first question is if companies are unable to obtain this
sensitive information from candidates, how can and should they
mitigate bias?
Ms. AJUNWA. So that's an excellent question. Of course, you
can't compel applicants to release information that's protected
information. However, employers can do analysis after the fact
to see if there is indeed a disparate impact based on looking
at for example the--this is after the fact, not during the
employment decision.
You know, just to look at the categories of people that
have applied versus the categories of people that were hired.
And this can then help them take steps in the future perhaps to
broaden their advertisement pool to attract more people from
protected categories if they are lacking those types of people.
Mr. TAKANO. A post-hiring review. Ms. Yang, while many
companies or vendors will claim that they are complying with
the EEOC regulations, we know that many do not because they are
currently, they currently operate in a gray area. Does the EEOC
have the ability to regulate the companies and vendors that are
contracted by employers to conduct hiring?
Ms. YANG. That's a very important question. The EEOC does
have an important role to play. The agency has sub-regulatory
authority under statutes like Title VII and can provide
guidance which is the Uniform Guidelines is one form of
guidance on how the agency believes vendors should validate
hiring screens.
So certainly, the agency could provide more up-to-date
guidance on some of the difficult issues where there are gray
areas.
You know, a lot of people say are correlations sufficient
to demonstrate that validity? I don't believe they are. It
would be helpful for the agency to make that clear and explain
why.
Mr. TAKANO. Well, what recourse if any does the EEOC have
to hold these companies accountable in the gray area?
Ms. YANG. Well, during the course of an investigation, I
mentioned earlier the EEOC has the authority to open its own
investigation on the commissioner's charge or directed
investigation depending on the statute.
So even if an individual doesn't have enough information to
come forward but the EEOC learns of a problem, it can open an
investigation. If it finds a problem, it can actually litigate
it to enforce the law.
But the challenge is having enough information to know
where the problems are because as you mentioned there is a very
big gap in knowledge because of the lack of transparency about
how many of these systems work.
Mr. TAKANO. Well, thank you. Thank you. Professor Ajunwa,
we know that auditing the algorithm and the code can help us
understand if the code is biased. But what kind of auditing
should be done and should it be the responsibility of the EEOC
to do this?
Ms. AJUNWA. Thank you very much for your question. So the
question of how the audits of automated hiring systems should
be completed or performed is one that I address in my two law
review articles, ``The Paradox of Automation as an Anti-Bias
Intervention'' and ``Automated Employment Discrimination.''
I don't come down on one side whether it has to be a
governmental agency, or it can be a third-party agency similar
to for example LEED, which certifies green buildings. So of
course, there is some utility in having it be a governmental
agency but also there is also the recognition of scarce
resources.
Mr. TAKANO. Well, so maybe, maybe not the government but
what kind of auditing should be done?
Ms. AJUNWA. So the kinds of auditing that should be done
should be one that's done with an, essentially an
interdisciplinary team so it should include lawyers, so labor
and employment lawyers.
It should include data scientists who are trained to write
code and to understand how machine-learning codes work. It
should include people who are versed in diversity research in
terms of creating a diverse workforce. So it should be an
interdisciplinary team.
Mr. TAKANO. Well, thank you. Madam Chair, while we should
not fear technology and the wonders of using it to increase
productivity and efficiency, we cannot move toward a society
where everything from employment to housing are guided by
systems that are largely unchecked. Thank you and I yield back
the balance of my time.
Chairwoman BONAMICI. Thank you, Mr. Takano. And finally,
last but not least, we will recognize Ms. Blunt Rochester from
Delaware for 5 minutes for your questions.
Ms. BLUNT ROCHESTER. Yeah. Thank you, Madam Chairwoman, for
this very important hearing and thank you to the panelists.
I had the opportunity last month, about two months ago to
start a future work caucus here in the Congress, a bipartisan
future work caucus and what you have shared today really
highlights the clarion call for all Members of Congress to be
engaged in this, in these discussions.
To me it appears that technology has really outpaced policy
and the people and so your participation here today is really
important. And there are so many topics, I wish I could have
had everybody's time that is not sitting here because there are
issues like language barriers that we haven't talked about, the
diversity of those people doing the design work, and making
sure that those algorithms are working and even returning
citizens.
I have a criminal justice bill called Clean Slate that
deals with people who are coming out of prison but therefore
are having challenges getting to work.
And I want to start off by finalizing Ms. Lee's question
because you two didn't get a, Ms. Landers or Mr. Romer-Friedman
didn't get a chance to answer the question about your one
concern, your big concern.
And then I want to ask the whole panel if there was one
thing Congress could do right now that would it be? So if I
could start with Ms. Landers and I have 3 minutes and 40
seconds left.
Ms. LANDER. So I'll be quick. The thing that occurred to me
is that we are really moving in a really positive direction in
terms of teleworking and worker flexibility.
However, there are a lot of laws like for example in
California and even under the FLSA that put such restrictions
on the employer and having to monitor very carefully--
Ms. BLUNT ROCHESTER. Yeah. So laws focusing on--
Ms. LANDER.--individual worktime that they're reluctant to
allow people the flexibility to telework. And so I think that's
a growing area because the generation that's coming up after me
really enjoys working from Starbucks.
Ms. BLUNT ROCHESTER. Yeah. So telework. Thank you. Mr.
Romer-Friedman.
Mr. ROMER-FRIEDMAN. Thanks. If I could just say this whole
line of argument that independent contracting is flexibility
and having employer employee relationship is not is a farce.
You can do all the flexibility you want through the
traditional employer relationship, get all the protections that
the New Deal and Great Society and subsequent laws created.
To your question, Congresswoman, I think that, you know, we
saw a scandal a couple weeks ago with Clearview that a company
essentially did, collected all photographs from people, from
pretty much everyone on the internet who was on social media
and created facial recognition, gave that to law enforcement
mostly.
I think it's concerning to me that we are seeing the
millennials and the next generation grow up in a time of social
media.
I think at some point employers will be able to literally
press a button and get every--from every piece of information
about someone that has been on the internet forever which not
just could be embarrassing to people but won't be
representative.
And if you point out, you know, someone who is returning
from prison from, who's, you know, who has paid their debt to
society with time, that person may have all that stuff come up
in the same way that right now you don't want a criminal record
even if it's not a conviction to be used for employment. So
thats going to be a big issue.
Ms. BLUNT ROCHESTER. Yeah. A big issue, thank you. And Dr.
Ajunwa, in terms of the thing that the Congresswoman--
Ms. AJUNWA. Yeah, thank you so much for your question. I
think really what is urgent for government to do right now is
to ensure that automated hiring is regulated. As it stands
there are just no regulations as to what information is
collected, how that information is evaluated and also what even
happens to that information whether the applicant is hired or
not.
And so governmental action is definitely needed both to
audit and certify the automated hiring system but also to
ensure that all the data that is being collected on applicants
is not something that's then used against the applicant in the
future.
Ms. BLUNT ROCHESTER. Yeah. Excellent, thank you. and, Dr.
Yang, Ms. Yang.
Ms. YANG. Thank you. I believe a workers' bill of rights is
needed so that workers understand how algorithms are making
decisions and how that might impact them.
Because people like Kyle Behm, he only knew he was screened
out because a friend who worked at the company told him. Most
people don't know this information and then the systems don't
get to improve from feedback loops about why people were
excluded, right.
So if people know, you know what, I don't think you're
going to accurately transcribe my accent with the type of
screen you're using, they can raise that concern and try to
make systems better.
Ms. BLUNT ROCHESTER. Yeah. Thank you. And I have so many
more questions which I will follow up with many of you
afterwards.
One question I did have for Dr. Ajunwa, you mentioned
microchips. I was just curious. I was looking for that in the
testimony. Could you speak a little bit more on that?
Ms. AJUNWA. Sure, thank you for your question. So your
question pertained to the use of microchips embedded under the
skin.
Ms. BLUNT ROCHESTER. Yeah.
Ms. AJUNWA. As a new really I think a surveillance tool. So
many corporations are marketing this as a convenience for
employees in terms of helping them to open doors or access
sensitive areas.
But in my opinion, because these chips are permanently with
the employee, they can track the employee wherever that
employee is, even off the job. So I do see them as surveillance
devices.
Ms. BLUNT ROCHESTER. Yeah. Well, my time has run out but I
thank you all for your testimony and we will be following up
with you.
I know data privacy is something that we are doing on my
other Committee, Energy and Commerce, so look forward to
working with you. Thank you and thank you, Madam Chairwoman.
Chairwoman BONAMICI. Thank you so much and I want to remind
my colleagues that pursuant to Committee practice, materials
for submission for the hearing record must be submitted to the
Committee Clerk within 14 days following the last day of the
hearing, preferably in Microsoft Word format.
The materials submitted must address the subject matter of
the hearing. Only a Member of the Committee or an invited
witness may submit materials for inclusion in the hearing
record. Documents are limited to 50 pages each.
Documents longer than 50 pages will be incorporated into
the record via an internet link that you must provide to the
Committee Clerk within the required timeframe, but please
recognize that years from now that link may no longer work. I
always like that part.
So again, I want to thank the witnesses for their
participation today. What we have heard is very valuable and I
think, I know many of us have a lot more questions and Members
of the Committee may submit those questions.
We ask that you please respond in writing. The hearing
record will be held open for 14 days to receive those
responses. And I remind my colleagues that pursuant to
Committee practice, witness questions for the hearing record
must be submitted to the Majority Committee staff or Committee
Clerk within 7 days and they must address the subject matter of
the hearing.
And I now recognize the distinguished Ranking Member for
his closing statement.
Mr. COMER. Thank you. Madam Chair, to begin with, I ask
unanimous consent to place in the record a statement from HR
Policy Association providing views on today's hearing topic.
Chairwoman BONAMICI. Without objection.
Mr. COMER. And I just again want to thank the witnesses who
came here to testify. This is an issue we are going to hear a
lot more about and we certainly want to be on top of this.
I appreciate all of the suggestions and like to often
remind this Committee we have a lot of laws already on the
books that address most of the subjects and topics that we
discuss in this Committee but it is always good to review the
issues as they emerge and make sure that if there is anything
that we can do in a bipartisan way in Congress to improve the
civil rights of workers then we certainly need to do that and
that is certainly a bipartisan issue.
But again, thank you all for being here today and, Madam
Chair, I yield back.
Chairwoman BONAMICI. Thank you very much Mr. Comer, and I
recognize myself for a purpose of making a closing statement.
And thank you again to your, to the witnesses for your
expertise which I very much appreciate.
I just want to reiterate that what the Ranking Member said
is that we often work on a bipartisan basis and I certainly
think that this is an issue where we could do that.
I know, Ms. Lander, you--a couple of times said things like
done correctly or used properly and I think those are the key
questions of the use of this technology if it is done correctly
and used properly, and I think that is the if that we are going
to be working on solving.
Today's hearing exposed how digital hiring, evaluation, and
management tools can threaten to replace civil rights
protections and left unchecked, these largely non-transparent
technologies can amplify and perpetuate existing biases that
intentionally or unintentionally discriminate against workers.
Our civil rights enforcement institutions and the laws they
enforce have not kept pace with the technologies that employers
are using to recruit, screen, interview and manage workers.
And as our modern workplaces continue to change and
employers increasingly rely on independent contractors, whether
misclassified or not, accountability for violations of workers'
basic civil rights can be diffused, and far too often many
workers will be excluded from key antidiscrimination
protections.
So Congress must fulfill its responsibility to preserve and
expand workers' civil rights by requiring transparency in the
algorithms that are used to recruit, hire, and evaluate
workers.
Preventing employers from stripping workers of their
antidiscrimination protections through misclassification and
clarifying, updating, and better enforcing our landmark civil
rights laws to meet the challenges workers face in the digital
age.
Technology has a tremendous amount of promise, but it is
the if used properly, if used correctly.
Congress has an opportunity to incentivize innovation in
workplace technologies that will put workers first and protect
and uphold equal employment opportunities.
And if we work together, we can shape a future in which
businesses can and will continue to innovate and workers can
and will enjoy strong antidiscrimination protections and I
think simply put: the future of work will be what we make it.
So thank you again. If there is no further business,
without objection, the Committee stands adjourned.
[Additional submissions by Chairwoman Bonamici follow:]
[GRAPHIC(S) NOT AVAILABLE IN TIFF FORMAT]
Help Wanted: An Examination of Hiring Algorithms, Equity,
and Bias: https://www.upturn.org/reports/2018/hiring-
algorithms/
[Additional submission by Mr. Comer follow:]
[GRAPHIC(S) NOT AVAILABLE IN TIFF FORMAT]
[Questions submitted for the record and their responses
follow:]
[GRAPHIC(S) NOT AVAILABLE IN TIFF FORMAT]
[Whereupon, at 3:34 p.m., the subcommittee was adjourned.]