Urging the Biden Administration to Address Technology’s Role in Hiring Discrimination

Harlan Yu, Aaron Rieke, and Natasha Duarte

Letter
Table of Contents keyboard_arrow_down

Upturn, ACLU, The Leadership Conference on Civil and Human Rights, and a coalition of other organizations urge the White House Office of Science & Technology Policy to fully incorporate the Biden administration’s commitment to racial equity into its AI and technology priorities.

This is our memo on technology's role in hiring discrimination. We sent this memo to:

  • Chair Charlotte A. Burrows, Equal Employment Opportunity Commission (EEOC)

  • Secretary Marty J. Walsh, U.S. Department of Labor (DOL)

  • Director Jenny R. Yang, Office of Federal Contract Compliance Programs (OFCCP), U.S. Department of Labor

  • Assistant Attorney General Kristen Clarke, Civil Rights Division, Department of Justice (DOJ)


Upturn was joined by the following signatories and contributors to this letter:

  • American Association of People with Disabilities

  • American Civil Liberties Union

  • Bazelon Center for Mental Health Law

  • Center for Democracy & Technology

  • Center on Privacy & Technology at Georgetown Law

  • Lawyers’ Committee for Civil Rights Under Law

  • The Leadership Conference on Civil and Human Rights


This is one of three memos on technology and discrimination that we sent to the Biden administration — see our letter to the White House Office of Science & Technology Policy; and our other two memos on housing and financial services discrimination.


RE: Addressing Technology’s Role in Hiring Discrimination

Too often, technology amplifies and exacerbates racial, gender, disability, economic, and intersectional inequity in our society. Governments and corporations, at the national, state, and local level, are using computer software, statistical models, assessment instruments, and other tools to make important decisions in areas such as employment, health, credit, housing, immigration, and the criminal legal system. In light of these developments, policymakers must take steps to clarify and strengthen guardrails to ensure non-discriminatory and equitable outcomes for all people seeking employment.

We offer the following proposals for the Biden-Harris administration for addressing discrimination arising from the use of new technologies in the hiring process. We urge all agencies to engage with a diverse range of stakeholders, including civil rights organizations, labor and workers’ rights organizations, and impacted communities, in order to receive ongoing input and feedback on these important issues. We also encourage agencies to prioritize transparency, both by sharing their data, models, decisions, and proposed solutions so that all stakeholders can stay apprised of and comment on the potential impact of proposed actions, and by encouraging employers and employment technology vendors to share with the public as much information as possible regarding their hiring technologies and assessments of those technologies. We would be pleased to discuss the ideas in this memo in more detail in the weeks and months ahead.

1. Gather information about employers’ use of hiring technologies, and proactively investigate discriminatory practices.

A. All agencies and departments should use their authority to collect information about the adoption, design, and impact of hiring technologies.

The EEOC has acknowledged that it needs information beyond individual complaints to direct its enforcement activities. Non-governmental researchers have stressed there are strict limits to non-privileged investigatory efforts. Fortunately, the Commission has the authority to “make such technical studies as are appropriate to effectuate the purposes and policies of [Title VII] and to make the results of such studies available to the public.”

However, to date, the Commission has not used these powers to gather information about employers’ adoption of hiring technologies. The Commission should use its research and data gathering authority to help guide enforcement efforts at all levels of government and spur the development of new industry standards. When appropriate, these studies could be conducted collaboratively with large employers or technology vendors and in consultation with the Equitable Data Working Group established under Executive Order 13985.

Other agencies and offices, such as the Bureau of Labor Statistics, the Women’s Bureau, and the Office of the Secretary of Labor, have broad statutory grants of authority to gather data and conduct investigations relating to the workings of the labor market. In some instances, these grants of authority are not being utilized, or at least not being utilized to the degree they could be. Given the degree to which employers and vendors have an information advantage in this space, agencies should be proactive and creative in their strategies to collect data and gain glimpses into the nature and extent of employers’ use of hiring technologies.

B. The Office of Federal Contract Compliance Programs (OFCCP) should strengthen its audits and rules.

The OFCCP has a unique and important role to play in guarding against discrimination. Federal contractors must take affirmative action to “ensure that applicants are employed . . . without regard to their race, color, religion, sex, sexual orientation, gender identity, or national origin,” and to “employ and advance in employment qualified individuals with disabilities.”

To help ensure employers are meeting these obligations, the OFCCP selects several thousand employers for a routine audit each year. However, historically, desk audits have been somewhat coarse, and there is evidence that employers’ compliance with recordkeeping requirements could be more robust. For example, the supporting documentation required for a desk audit does not include any information about the selection procedures, technologies, or criteria employers use. Moreover, according to a 2020 report, the OFCCP observed that many employers failed to document the number of applicants and hires identifying as people with disabilities.

The OFCCP should consider expanding and strengthening its auditing regime, particularly in light of the fact that large employers can easily collect and evaluate detailed records. To this end, the OFCCP could expand the supporting documentation required for a routine desk audit to include the specific selection procedures the employer uses for each job or job group, the employers’ processes or methodologies for evaluating the selection procedures’ adverse impacts, and any alternative selection procedures the employer has designed or provided for applicants with disabilities.

Moreover, the OFCCP should leverage its access to employer data to help the public better understand the hiring technologies used by federal contractors. Federal contractors hold extensive records in modern applicant tracking systems regarding their hiring practices and outcomes, including progress toward affirmative action goals and the impacts of hiring tests and technologies. But the scheduling letter that OFCCP sends to thousands of employers each year does not request any specific information about employee selection devices. The OFCCP should strengthen employers’ reporting requirements and, to the extent permitted by law, make more information public regarding the nature, frequency, and legality of hiring practices that it encounters during the course of its enforcement activity.

Finally, the rapid adoption of new hiring technologies adds urgency to federal contractors’ obligation to identify opportunities in agency policies, regulations, and guidance to address systemic inequities for people with disabilities. Beyond enhanced auditing and data requirements, the OFCCP should revisit its regulations under Section 503 of the Rehabilitation Act of 1973, and align them with EEOC’s Section 501 regulations. Today, Section 503 rules provide lower expectations for federal contractors’ affirmative action efforts to employ people with disabilities than parallel rules under Section 501, which require the same of federal agencies. The more ambitious goals of Section 501, including a subgoal for people with targeted disabilities, should be imposed for Section 503 as well.

Job applicants “typically lack information about a discriminatory hiring policy or practice,” making it difficult for them to file complaints. Hiring is only becoming more opaque as employers rapidly adopt new hiring technologies. As of January 2021, the Commission reported that it had not acted upon a single complaint involving the use of new hiring technologies. It is clear a more proactive approach is needed.

When supported by its research and information gathering, the EEOC should consider use of Commissioner charges under Title VII and the Americans with Disabilities Act (ADA), and directed investigations under the Age Discrimination in Employment Act (ADEA) and the Equal Pay Act, to investigate discrimination in the absence of individual complaints. Tackling systemic discrimination is “central to the mission of the EEOC,” and new technologies at all stages of the hiring process raise systemic concerns. 

2. Modernize guidance for selection procedures and expand recordkeeping requirements.

A. The DOL, DOJ, and EEOC should begin a process to update the Uniform Guidelines on Employee Selection Procedures (UGESP).

For many employers and technology vendors, the Uniform Guidelines on Employee Selection Procedures (UGESP) are a key source of guidance for assessing the fairness and validity of hiring selection procedures. Unfortunately, the UGESP, which were adopted in 1978, are showing their age. They were adopted before Congress passed the Americans with Disabilities Act, and they do not address discrimination against people with disabilities, age discrimination, the full scope of sex discrimination, or intersectional discrimination. They lag far behind modern scientific standards, such as the American Psychological Association’s Principles for the Validation and Use of Personnel Selection Procedures (which acknowledges that fairness is an integral part of the validity analysis). And there is a risk that they could be interpreted to allow employers to establish that a hiring procedure with a disparate impact is “valid” based on correlational evidence alone — even if the correlation is based on factors clearly unrelated to job performance.

In the short term, the EEOC should emphasize the UGESP’s limited application as a standard for anti-discriminatory hiring practices, including through a revised and updated “Question and Answers” resources. The EEOC should also publish informal guidance encouraging hiring technology vendors and employers to design, test, and audit their technologies to prevent discriminatory effects. Such guidance would not only provide more clarity for employers as to the kind of information that they should be seeking from vendors when procuring hiring technologies and the testing they themselves need to conduct while using them, but also would encourage the creation of third-party auditing standards. Ultimately, however, formal updates to the UGESP are needed to cover recruitment and sourcing practices, create standards for non-statistical auditing, address disability discrimination, and further elevate the importance of job-relatedness. Federal agencies should begin a multistakeholder process to update the UGESP for the modern era.

B. All agencies responsible for enforcement of employment discrimination laws should adopt OFCCP’s Internet Applicant Rule (IAR), and clarify that it applies to all persons assessed and screened by hiring technologies.

The OFCCP’s IAR has long been the most influential standard for determining when an individual becomes an “applicant” for purposes of recordkeeping and compliance with antidiscrimination laws. But several factors have limited the IAR’s practical reach in recent years. First, the IAR only considers someone an “applicant” if they make “an expression of interest in employment” that “indicate[s] the individual possesses the basic qualifications for the position.” But companies’ sourcing efforts have increasingly relied on web-based advertising and searches of databases containing “passive” candidates who never expressed interest in a particular job or employer, and whose profiles may not indicate their qualifications. It is not clear whether such passive candidates are “applicants” under the IAR. 

Second, a candidate is no longer considered an applicant under the IAR if the candidate removes themselves from consideration for any reason, or if they are screened out by automated processes without even having an opportunity to “apply.” This places a burden on disabled workers, many of whom may be unable to continue with Internet-based application processes that are inaccessible to them if the employer does not offer an accommodation that would allow them to continue to pursue an employment opportunity further. Finally, no agency other than OFCCP has formally endorsed or adopted the IAR, which reduces the rule’s effective scope as the use of such sourcing methods spreads across the labor market.

To ensure that recordkeeping obligations and compliance efforts track the reality of modern sourcing and recruitment methods, OFCCP should clarify that the IAR applies to every person screened by hiring technologies — even if they never formally applied for a specific position. The IAR should further specify that if a person removes themselves from consideration due to inaccessibility or lack of accommodation, that person should still be considered an applicant. The EEOC, DOJ, and all other agencies responsible for investigating and enforcing employment discrimination laws should then formally adopt the updated rule.

C. All agencies with responsibilities under the ADA and Section 504 of the Rehabilitation Act should endorse new guidance for diversity and inclusion efforts for people with disabilities.

The Office of Disability Employment Policy (ODEP) is crucial for shaping how procurement and development processes for hiring technologies will impact jobseekers with disabilities. In July 2020, EARN and PEAT developed the Checklist for Employers: Facilitating the Hiring of People with Disabilities Through the Use of eRecruiting Screening Systems, Including AI through an ODEP grant. This checklist poses very useful questions for a range of stakeholders involved in the procurement and development process, but they should be expanded to explain these best practices in greater detail to facilitate their adoption by employers. Federal disability law protects job applicants from being forced to disclose disability prior to a conditional job offer in order to fairly participate in the hiring process. This might require making alternative evaluation methods readily available to all applicants at the outset rather than only by request, and proactively addressing ways in which the technologies increase the likelihood of disclosure. All federal agencies responsible for enforcement and/or implementation of the ADA or Section 504 should endorse the checklist’s open-ended questions about diversity and inclusion efforts, assessment design, testing, and auditing. 

D. The EEOC should require employers to retain records about reasonable accommodations and applicants with disabilities.

The EEOC should require employers to maintain records about the types of reasonable accommodations offered when hiring technologies are used. Records should indicate when and how reasonable accommodations are offered, requested, and denied, as well as the incidence of job rejections under each situation. They should also confirm a robust and transparent process was utilized. Employers should proactively provide for a range of reasonable accommodations, even without request, so as to expand the field of people with disabilities who can access employment opportunities. Guidance should acknowledge that tools such as resume mining software will likely require human review to ensure that candidates are being fairly assessed based on job-related skills and qualifications. The EEOC should establish a process for periodically collecting this information from employers.

The EEOC and the Office of Management and Budget (OMB) should also consider measures to include the collection of data on disability status and targeted disability status in the EEO-1 data collection process. This data is essential to fully assessing employers’ compliance with relevant civil rights laws.

3. Hold workshops and convenings on hiring technologies, and make more information available to the public.

All relevant agencies should hold workshops and convenings to gather more information about industry practices. The EEOC held a meeting in 2016 on the implications of big data for equal employment opportunity. However, it did not release a report or public guidance to share any lessons or outcomes of the meeting. All agencies should use their convening authority to gather industry, academics, advocates, impacted communities and other experts to share more information with the public about best practices and enforcement activity.

Thank you for your attention to these matters. For any questions or further discussion, please contact Aaron Rieke, Managing Director, Upturn, at 202-677-2359 or aaron@upturn.org.

Download


1

EEOC, Advancing Opportunity: A Review of the Systemic Program of the U.S. EEOC (July 7, 2016), https://www.eeoc.gov/advancing-opportunity-review-systemic-program-us-equal-employment-opportunity-commission (“[I]t is important for EEOC to consider additional sources of information in identifying issues where government enforcement is most needed . . . For example, EEOC’s Research and Data Plan identifies the need for research on screening devices, tests, and other practices to identify barriers to opportunity as well as promising selection practices that rely on job-related criteria.”).

1

See, e.g., Upturn, Essential Work: Analyzing the Hiring Technologies of Large Hourly Employers, May 2021, https://www.upturn.org/reports/2021/essential-work/.

1

Section 705(g)(5) of Title VII. Similar authority is granted by the ADA. (42 U.S.C. § 12117(a)).

1

EEOC Letter to Senator Bennet, January 15, 2021, available at https://www.upturn.org/static/files/2021-01-15-EEOC-Bennet-Response.pdf.

1

Executive Order 13985, Advancing Racial Equity and Support for Underserved Communities Through the Federal Government, January 20, 2021, available at https://www.whitehouse.gov/briefing-room/presidential-actions/2021/01/20/executive-order-advancing-racial-equity-and-support-for-underserved-communities-through-the-federal-government/.

1

See, e.g., 29 U.S.C. § 2 (charging Bureau of Labor Statistics with the duty to “investigate the causes of, and facts relating to, all controversies and disputes between employers and employees as they may occur, and which may tend to interfere with the welfare of the people of the different States”); 29 U.S.C. § 13 (granting Women’s Bureau “authority to investigate and report to the Department of Labor upon all matters pertaining to the welfare of women in industry”); 29 U.S.C. § 9 (granting the Secretary of Labor authority “to make special statistical studies relating to employment, hours of work, wages, and other conditions of employment”).

1

EO 11246 § 202.

1

29 U.S.C. § 793.

1

OFCCP, Corporate Scheduling Announcement List (CSAL) Frequently Asked Questions, available at https://www.dol.gov/agencies/ofccp/faqs/scheduling-lists.

1

OFCCP, Federal Contract Compliance Manual, Chapter 1 Desk Audit, available at https://www.dol.gov/agencies/ofccp/manual/fccm/chapter-1-desk-audit.

1

1

A robust auditing regime would align with the goals of Executive Order 13985 to promote equitable delivery of benefits and equitable opportunities.

1

OFCCP, Compliance Check Scheduling Letter, available at https://www.dol.gov/agencies/ofccp/compliance-checks/scheduling-letter.

1

EEOC, Advancing Opportunity A Review of the Systemic Program of the U.S. Equal Employment Opportunity Commission, July 7 , 2016, https://www.eeoc.gov/advancing-opportunity-review-systemic-program-us-equal-employment-opportunity-commission.

1

EEOC Letter to Senator Bennet, January 15, 2021, available at https://www.upturn.org/static/files/2021-01-15-EEOC-Bennet-Response.pdf.

1

EEOC, Advancing Opportunity A Review of the Systemic Program of the U.S. Equal Employment Opportunity Commission, July 7 , 2016, https://www.eeoc.gov/advancing-opportunity-review-systemic-program-us-equal-employment-opportunity-commission.

1

See, e.g., Upturn, Help Wanted: An Examination of Hiring Algorithms, Equity, and Bias, December 2018, https://www.upturn.org/reports/2018/hiring-algorithms/; Cener for Democracy & Technology, Algorithm-driven Hiring Tools: Innovative Recruitment or Expedited Disability Discrimination?, December 2020, https://cdt.org/insights/report-algorithm-driven-hiring-tools-innovative-recruitment-or-expedited-disability-discrimination/.

1

American Psychological Association, Principles for the Validation and Use of Personnel Selection Procedures, August 2018, available at https://www.apa.org/ed/accreditation/about/policies/personnel-selection-procedures.pdf.

1

In developing guidance, EEOC should review existing literature and frameworks such as Dillon Reisman et al., AI Now, Algorithmic Impact Assessments: A Practical Framework for Public Agency Accountability (Apr. 2018), https://ainowinstitute.org/aiareport2018.pdf; Brookings Institute, Auditing employment algorithms for discrimination, https://www.brookings.edu/research/auditing-employment-algorithms-for-discrimination/ Inioluwa Deborah Raji, et al., Closing the AI Accountability Gap: Defining an End-to-End Framework for Internal Algorithmic Auditing (2020); Timnit Gebru and Jamie Morgenstern, et al., Datasheets for Datasets (March 19, 2020); Margaret Mitchell and Simone Wu, et al., Model Cards for Model Reporting (Jan. 2019)Office of the Comptroller of the Currency, Supervisory Guidance on Model Risk Management, Apr. 2011, https://www.occ.gov/news-issuances/bulletins/2011/bulletin-2011-12a.pdf.

1

41 CFR § 60-1.3.

1

Employer Assistance and Resource Network on Disability Inclusion, Checklist for Employers: Facilitating the Hiring of People with Disabilities Through the Use of eRecruiting Screening Systems, Including AI, https://askearn.org/wp-content/uploads/2020/07/EARN_PEAT_eRecruiting_Checklist.pdf.

1

EEOC, Big Data in the Workplace: Examining Implications for Equal Employment Opportunity Law, October 13, 2016, https://www.eeoc.gov/meetings/meeting-october-13-2016-big-data-workplace-examining-implications-equal-employment.