Workplace AI Under the Microscope: Shifting Regulatory Landscape and Audit Implications
The U.S. regulatory landscape for AI in the workplace is rapidly evolving, with a significant retreat of federal guidance and a surge in state-level legislation and private litigation. Internal audit and assurance professionals must understand this shift to effectively assess and mitigate risks associated with AI deployment in HR and operational processes, particularly concerning anti-discrimination laws and potential vendor liabilities.
Federal Retreat Creates Regulatory Vacuum
The federal government, particularly the Equal Employment Opportunity Commission (EEOC), has significantly scaled back its guidance and enforcement efforts regarding artificial intelligence in the workplace. Executive orders have deprioritized disparate-impact liability, and the EEOC has removed key AI guidance documents from its website. This federal retreat, however, does not eliminate existing anti-discrimination statutes like Title VII, the Americans with Disabilities Act, and the Age Discrimination in Employment Act. Instead, it removes federal safe harbors and official interpretations, leaving employers with statutory obligations but less clear federal guidance, thereby increasing uncertainty and exposure.
States and Courts Step Up Enforcement
Into the vacuum left by federal agencies, state legislatures and the plaintiffs' bar have emerged as the primary drivers of AI governance in the workplace. States like Illinois, Colorado, and New York City have enacted new laws or strengthened existing ones to address discriminatory AI use, mandate impact assessments, and require bias audits. This creates a complex, fragmented regulatory environment where employers operating across state lines must adhere to the strictest applicable standards. The shift means that the courtroom, through private litigation and class-action lawsuits, and statehouses are now the central arenas for defining and enforcing responsible AI practices in employment.
Key Cases Highlight Enduring Liabilities
The article highlights the enduring nature of liability for discriminatory AI, even in the absence of federal guidance. The case of EEOC v. iTutorGroup demonstrated that technology-enabled discrimination, regardless of the AI system's complexity, leads to significant penalties. This precedent remains relevant, as the underlying anti-discrimination statutes are still in force. Furthermore, the ongoing Mobley v. Workday litigation is reshaping vendor liability for AI tools, extending state laws to applicants nationwide and revealing how bias-testing evidence can be shielded under privilege. These cases underscore that employers remain accountable for the outcomes of their AI systems, and the responsibility for ensuring fairness and compliance now largely rests with internal legal teams, state regulators, and private plaintiffs.
Read more
