The hiring process is growing increasingly automated, as employers in New Jersey and around the country turn to artificial intelligence (AI) that uses hiring algorithms. This could be a time-saving measure for employers, helping them sort through large numbers of job applications, but it can also potentially result in violation of antidiscrimination laws. While it might seem unlikely that employers would use these algorithms for deliberate discrimination against categories protected by laws like the New Jersey Law Against Discrimination, not all unlawful employment discrimination requires discriminatory intent. AI relies on the information it receives from human users. When an AI bases hiring recommendations on existing data or past hiring patterns, it could end up perpetuating inequities. One question that courts are only beginning to address that could have an impact on New Jersey employment discrimination cases is how to determine liability when a computer engages in discrimination on an employer’s behalf.
Disparate Impact Discrimination
A policy or practice with no discriminatory intent can still violate antidiscrimination laws if it has a disparate impact on members of a protected group when compared to others. The U.S. Supreme Court first recognized disparate impact discrimination in Griggs v. Duke Power Co., 401 U.S. 424 (1971). The employer, located in North Carolina, required applicants for certain jobs to have a high school diploma and to pass an IQ test. The policy had a substantially disparate impact on Black applicants. The court found that the employer’s policy was not “reasonably related” to the jobs in question, and that it therefore violated Title VII of the Civil Rights Act of 1964.
Discrimination in, Discrimination Out
AI algorithms are designed to simulate the functions of the human brain. They cannot “think” on their own — at least not yet. Instead, they process information according to algorithms, all provided by humans. Hiring AIs consider metrics like education and work experience. The most sophisticated hiring tools available right now can even analyze video recordings of job interviews to evaluate candidates.
The term “garbage in garbage out” originated in the early days of computing, and it is relevant to this subject. If the information or instructions provided to a hiring AI are biased against one or more protected groups, even unintentionally, the hiring AI will almost certainly produce biased results. A hiring AI developed by a major tech company, for example, reviewed resumes submitted over the prior ten years in order to identify patterns. This resulted in hiring recommendations that favored men over women because of existing gender imbalances in the tech industry. The company abandoned the system as a result.
Regulatory and Court Challenges to AI Discrimination
Few courts have addressed this kind of discrimination so far, but regulatory agencies have taken note. The Equal Employment Opportunity Commission held a hearing on the issue in 2016. Liability is likely to be a question in discrimination lawsuits. An aggrieved job applicant can only file suit against the employer. The employer could then bring the AI developer into the suit, alleging that they are to blame for the AI’s actions.
A group that advocates for online privacy rights filed a complaint last year with the Federal Trade Commission alleging deceptive trade practices by a company that develops technology to scan and analyze job applicants during video interviews. The complaint does not allege violations of employment laws, but it could have an impact on future employment discrimination claims.
The Resnick Law Group’s employment attorneys are available to help you if you have a gender discrimination or other dispute with your employer in New Jersey or New York. Please contact us today at 973-781-1204, at 646-867-7997, or online to schedule a confidential consultation to discuss your case.