AI is changing the way we live, work and interact with one another. Disruptive technologies are fundamentally altering our world with advances in artificial intelligence, machine learning, big data analytics, robotics, cloud technology, autonomous vehicles, biotechnology and virtual and augmented reality. These developments are bringing about a 4th Industrial Revolution, a phrase coined by Klaus Schwab, founder and executive chairman of the World Economic Forum, who authored an influential book on the subject.
This revolution impacts almost every aspect of our lives, including the workplace. Technology runs ahead of the both government’s ability to create common sense legislation that would rein in technology run amok, and the ability of some businesses to swiftly adapt and transition to the new normal. Human Resource departments of businesses large and small are navigating an environment that’s murky and ever-evolving.
For our three-part series on AI In the Workplace, we’re turning to regional professionals who are subject matter experts in the fields of law, technology and Human Resources to provide some direction. In this edition, Milena Radovic, an associate in the law office of Jackson Lewis P.C. answers our questions. Her practice focuses exclusively on workplace law, including employment litigation, workplace training, and preventative client counseling.
CoVaBIZ: How are employers currently using AI technology in the realm of human resources?
Radovic: Employers are using AI in a multitude of ways to maximize their efficiency, eliminate redundancies, and improve workplace safety. For example, companies are using AI in the recruitment and hiring process by conducting candidate searches, reviewing and screening applicant resumes, selecting applicants for interviews, reviewing interview response, and scoring applicants on skills assessments. Human Resources Departments can use AI to rewrite policies and generate personnel forms.
Likewise, AI can provide input into the drafting of disciplinary action notices and performance evaluations. Employers can also utilize AI to make difficult decisions regarding reasonable accommodations, pay determinations, selections for promotion, and ultimately separations from employment. Companies can use AI to minimize accidents in the workplace by monitoring for fatigue, fall detecting and workplace safety accidents and identifying and evaluating hazardous conditions and exposure risks.
CoVaBIZ: AI can increase efficiencies, but we’re seeing examples of how AI productivity tools used to hire and monitor workers puts some at a disadvantage. In the recruiting process, AI can’t determine why gaps in employment may have occurred.
Radovic: Yes, With possibilities come potential pitfalls. Interestingly, AI advocates argue that using AI can help eliminate human bias and subjectivity. While this may be advantageous in theory, we have seen real world applications where AI actively showed bias against persons of color, women, and persons with disabilities – despite this being the very thing that it is supposed to prevent against. From a legal standpoint, one of the many concerns of using AI to make employment decisions is that it may have a disparate impact on employees with protected characteristics.
For example, using AI in the recruitment may result in persons with gaps in their employment being screened out, but a gap in employment may be the result of the applicant having experienced a medical condition which necessitated an extended leave of absence or a woman temporarily stepping away from the workforce after the birth of a child. In such situations, companies may find themselves unintentionally violating federal and state anti-discrimination statutes, such as Title VII of the Civil Rights Act of 1964, Americans with Disabilities Act of 1990, Virginia Human Rights Act, and the Virginians with Disabilities Act.
CoVaBIZ: What kind of guidance is available to employers and HR managers relating to AI and employment decisions?
Radovic: In recent years, the U.S. Equal Employment Opportunity Commission (“EEOC”) has issued two separate guidance documents related to AI. In 2022, it issued “The Americans with Disabilities Act and the Use of Software, Algorithms, and Artificial Intelligence to Assess Job Applicants and Employees” and earlier this year, it issued “Assessing Adverse Impact in Software, Algorithms, and Artificial Intelligence Used in Employment Selection Procedures Under Title VII of the Civil Rights Act of 1964.” This issue is not going away anytime soon. In fact, the EEOC included it in its “Draft Strategic Enforcement Plan for 2023—2027.”
Specifically, the EEOC is interested in eliminating barriers in recruitment and hiring, including the use of automated systems, including artificial intelligence or machine learning, to target job advertisements, recruit applicants, or make or assist in hiring decisions where such systems intentionally exclude or adversely impact protected group and screening tools or requirements that disproportionately impact workers based on their protected status, including those facilitated by artificial intelligence or other automated systems, pre-employment tests and background checks.
As such, employers should be cautious when using AI to make employment decisions. Simply relying on outside vendors and their promises of bias-free decision-making is not enough to ensure compliance with federal and state law.
CoVaBIZ: What about employers who operate in multiple states?
Radovic: Multi-state employers should also exercise caution. While Virginia has not currently enacted any legislation addressing use of AI in the workplace, other jurisdictions have. For example, New York City Local Law 144 went into effect July 5, 2023. In relevant party, NYC 144 prohibits employers and employment agencies from using an automated employment decision tool unless the tool has been subject to a bias audit within one year of the use of the tool, information about the bias audit is publicly available, and certain notices have been provided to employees or job candidates. When one jurisdiction passes legislation, regulations, ordinances, or guidance, other jurisdictions are sure to follow. As such, it’s important that multi-state employers stay up to date on current and pending legislation, regulations, ordinances, or guidance.
CoVaBIZ: It would seem that issues surrounding transparency, notice and consent can be problematic for employers. Are there best practices?
Radovic: While proposed legislation is expected to address these issues, currently there is limited guidance. Nonetheless, as a best practice, employers should consider whether they want to provide notice and consent, whether express or implied, about the company’s use of AI in the recruitment and recruitment process. Employers should also consider potential privacy concerns, including AI gathering employee data and how is that data being collected, used, stored, and shared. Some states, including Virginia, have passed legislation related to how consumer and employee data is collected and shared.
CoVaBIZ: Some employers and HR departments may be looking to AI to help them draft policy as a cost savings measure. Can this backfire?
Radovic: While AI can certainly serve as resource in helping HR departments, it’s not a substitute for outside legal review. For example, AI may help a company draft a reasonable accommodation policy, but the AI drafted policy may lack the nuances to be compliant with federal and state law. During the height of the pandemic, the Virginia General Assembly revised the Virginia Human Rights Act to provide reasonable accommodations to qualified employees.
One often overlooked aspect of the amended Virginia Human Rights Act is that covered employers must post in a conspicuous location and include in any employee handbook information concerning the employee’s right to a reasonable accommodation for disabilities and known limitations related to pregnancy, childbirth, or related medical conditions. Covered employers must also provide this information directly to the employee upon commencement of employment and within ten (10) days of the employee providing the employers with notice of their condition.
Ultimately, AI is a limited resource because it lacks the required background knowledge to offer information that is fully legally compliant. Although using AI to draft an employee handbook policy may initially seem to be an effective cost-cutting decision, it can have costly, unintended outcomes.