EEOC commissioner advises CHROs: Don’t let AI ‘scale discrimination’

The rights of employees must be protected when employment-related decisions are subjected to AI, Keith Sonderling, commissioner of the U.S. Equal Employment Opportunity Commission, recently told a virtual audience of HRE readers. In a conversation with Madeline Laurano of Aptitude Research, the commissioner cautioned HR leaders to maintain a principled approach when using technology to aid employment-related decisions.

- Advertisement -

HR leaders are uniquely positioned to spearhead AI governance within their organizations, Sonderling said, because practitioners are aware of civil rights and accustomed to operating within a highly regulated environment. He emphasized the impact of HR technology, citing products from “A to Z” of the employment relationship, encompassing everything from job descriptions to performance management and beyond.

Tech can increase discrimination

While artificial intelligence has the potential to broaden the applicant pool and eliminate persistent human bias, it also poses challenges, he says. Responsible HR principles must be built into the use of any tech in the same way they are applied to other functions of the field.

Keith Sonderling, EEOC; EEOC commissioner Keith Sonderling on AI and employee rights
Keith Sonderling, EEOC commissioner

For example, consider the power and speed made possible with machine learning, which will accelerate the reach of employment decisions. Sonderling says that companies might unintentionally use AI to replicate a “discriminatory status quo” by injecting bias into the algorithm, leaving AI to “scale discrimination to the likes that we’ve never seen.”

HR leaders must carefully consider the data that is trained into large language models, ensuring that machines do not replicate discriminatory practices. Today’s HR tech vendor landscape offers products that can assist in every aspect of candidate and employee lifespans, and Sonderling emphasizes that these platforms should focus on actual job-related capabilities while avoiding the revelation of protected characteristics such as national origin, disability, age and others.

“If you can’t account for [a person’s] characteristics, the tech can’t be used properly,” says Sonderling. This is because unexpected challenges arise when AI struggles to accurately perceive certain human traits, such as accents or communication styles. In such instances, HR leaders are responsible for ensuring technology is used judiciously to prevent discrimination.

AI pitfalls that violate equal opportunity

The pressure to implement AI is often mentioned alongside predictions of workforce reduction. HR leaders must navigate this transition meticulously to prevent discrimination by thoughtfully examining roles that will be eliminated, according to Sonderling.

- Advertisement -

The commissioner reminds HR teams to evaluate scenarios that could have a disparate impact on certain employee groups. Sweeping moves such as letting go of the highest-paid staff or terminating the most recent group of hires might produce results that unintentionally churn inequity.

Sonderling also warns of unseen equal opportunity pitfalls around the hot topic of skilling and reskilling. Older workers may feel they are “forced to quit” due to drastic changes in how they perform their jobs, while employees with disabilities might require adaptive devices or accommodations, he says. HR leaders must navigate such skill-related workforce changes while keeping current employment laws in mind.

Related: Learn more from Keith Sonderling and Madeline Laurano at HR Technology Europe later this year. Both will keynote the event, which will be held May 2-3 in Amsterdam. Register here.

Human resource leaders should also be aware of the mental health implications of implementing new technology. Sonderling says the EEOC has witnessed reports indicating anxiety and fear surrounding AI in the workplace, particularly gen AI. Concerns about job displacement resulting in PTSD, depression and anxiety may necessitate HR interventions and accommodations.

Organizations “can’t have a set it and forget it approach,” he says, reminding employers that AI tools must be “tailored” to ensure they are “doing what they are supposed to do.”

AI must be ‘carefully designed and properly used’

Madeline Laurano, Aptitude Research
Madeline Laurano

“No company wants to buy a product that violates an employee’s civil rights,” says Sonderling. However, all liability will fall on business leaders, not technology vendors or the raw output of artificial intelligence. He says that care, caution and governance are imperative. Laurano reminds HR leaders that tech-related investments should be made with “longstanding civil rights laws in mind.”

The rules and regulations of the EEOC aren’t placed on technology—they apply, as always, to U.S.-based employers and organizations. Stick to the fundamental principles of human resources, according to the commissioner, because ultimately, technology should facilitate—not be responsible for— HR processes.

Sonderling reminds business leaders that the future of HR and AI integration, according to the EEOC, will require balancing technological advancement with the preservation of employees’ rights within all existing laws. AI, he says, must be “carefully designed and properly used.”

Learn more: Hear the full conversation between Sonderling and Laurano on the webinar replay.

Jill Barth
Jill Barth is HR Tech Editor of Human Resource Executive. She is an award-winning journalist with bylines in Forbes, USA Today and other international publications. With a background in communications, media, B2B ecommerce and the workplace, she also served as a consultant with Gallagher Benefit Services for nearly a decade. Reach out at [email protected].