The ACLU recently issued a clear warning to employers, urging the FTC to investigate personality assessment tests, video interview tools, and cognitive ability assessment screening devices (all of which are powered by artificial intelligence) for possible discrimination. The May 30 complaint comes after the ACLU filed a complaint with the EEOC regarding the same AI tools, alleging that the tools unfairly screen out applicants with disabilities and unfairly target applicants from certain racial backgrounds. What do employers and AI developers need to know about these aggressive tactics, and how should employers adjust?
Three recruiting tools targeted by the ACLU
The three tools targeted in the attack were developed by Aon Consulting, a leading firm in the evaluation field.
The ACLU argues that these tools assess general personality traits that are not work-related or necessary for a specific job, such as proactiveness, emotional awareness, energy, ambition, and drive. They argue that people with depression, anxiety, and autism (to name just a few) may receive lower ratings regardless of their actual job skills, potentially unfairly screening people out based on their disability.
The ACLU specifically says that using AI in these tools only exacerbates the underlying problems: Indeed, they argue, video interview tools could discriminate based on disability, race, and other protected characteristics because the programs rely on AI data.
The ACLU cites data showing alarming disparities in the average Grid Challenge scores of test-takers of different races, alleging that studies have revealed that Asian, black, Hispanic, Latino and other test-takers received lower average scores than white test-takers.
They also claim that gridChallenge may discriminate against people with cognitive disabilities and various mental disorders – for example, they claim that people with autism score significantly lower on various measures of working memory compared to the general population.
Why did the ACLU file a lawsuit with the FTC?
The complaint the ACLU filed with the FTC centers around the developers' claims that these recruiting tools are “fair,” “without bias,” and “without adverse effects,” which the ACLU equates to deceptive marketing tactics. It alleges there are representations, omissions, or practices that could mislead and harm consumers (in this case, employers looking for an efficient and effective recruiting experience).
The groups are calling on the authorities to open an investigation into whether Aon's developers have engaged in unfair and deceptive acts and practices, and to order Aon to stop selling or managing these tools until those issues are resolved.
Employers are not necessarily immune from liability
The ACLU's complaint to the FTC doesn't name any employers who may have relied on these tools, but the ACLU hasn't completely forgiven them yet. While it argues that the developers may have misled employers with false statements about the tools, it also says that “employers have a legal obligation to ensure that the ratings they use are non-discriminatory, including when using technically complex ratings.” Plaintiffs' lawyers will likely be watching the FTC's investigation closely.
EEOC process also ongoing
In fact, the ACLU has already launched an employment-related attack against Aon and one employer that uses these tools. The ACLU alleges that a biracial, autistic job applicant who was asked to take the ADEPT-15 and gridChallenge assessments as part of the employer's hiring process faced racial and disability discrimination as a result. The ACLU alleges violations of the Americans with Disabilities Act (ADA) and Title 7 of the Civil Rights Act of 1964 against that applicant, as well as all other similarly situated job applicants.
What's next?
The EEOC charges were filed late last year and are still pending before the agency, and while the administrative process could take several more months before any action emerges, the charges are likely just a precursor to the ACLU filing a federal lawsuit against Aon and the employer with similar allegations.
The FTC complaint is quite recent, having only been filed in the past few weeks. Based on the allegations set out in the complaint and the developer's response, the FTC will soon determine what action to take, but the administrative process could take several months to complete.
What should employers do?
If you’re currently hiring AI developers or looking to improve your hiring process with AI-powered tools, you should take these recent actions into consideration and consider your next steps.
- Establish an AI governance system – Develop clear systems and procedures, including a human oversight component, before introducing AI into the workplace, consistent with the National Institute of Standards and Technology (NIST) guidelines for artificial intelligence (NIST AI RMF), and evaluate the processes once they are in place to ensure you are effectively managing your AI systems.
- Vetting vendors – As the ACLU points out in its complaint, employers can’t just sit back and blame AI vendors when problems arise. The EEOC said as much when it issued guidance on AI bias last year. Plaintiffs are increasingly looking to hold employers responsible for resulting discrimination, so be sure to ask the right questions (with the help of a lawyer) when reviewing and implementing products.
- Be transparent – Employers will need to be transparent with employees and candidates about the use of AI systems in the workplace. This is an increasingly common theme and may soon be mandated in states such as Colorado, California and New York.
- Provide accommodations where possible – Some states may soon require employers to allow employees and applicants to opt out of AI processes that determine their future at work. You should offer similar options. Options could include specialized equipment, alternative testing or exam formats, or the opportunity to work in a quiet environment.
- Matching Questions – When creating a set of questions to feed into your AI system and ask candidates, make sure the questions closely align with the job requirements for the position and don’t deviate too far.
