On December 3, a bipartisan group of legislators lead by Representative Suzanne Bonamici (D-OR) introduced legislation that would prohibit employers from relying solely on automated decision-making systems to make employment-related decisions. HR 6371, known as the No Robot Bosses Act, would mandate that employers include human oversight when they use automated decision-making software, that they regularly test their tools, and that they disclose to workers when such tools are being used.
What Would be Covered
The bill defines an “automated decision system” as one that “uses computation” to “determine outcomes” and “make or aid decisions” and includes those “derived from machine learning, statistics, or other data processing or artificial intelligence techniques.”
Any employment-related decision by an automated decision system would have to be independently corroborated by “meaningful oversight by a human with appropriate and relevant experience.” Employment-related decisions include actions relating to recruiting, screening, interviewing, or selecting a candidate; firing, retaining, taking disciplinary action against, demoting, or reassigning an employee; and implementing other actions affecting terms and conditions of employment such as compensation, scheduling, insurance coverage, or benefits.
Employer Verification and Disclosures
The bill would require employers to verify that any automated decision system complies with federal antidiscrimination laws. Employers would be required to analyze the results of their automated decision system annually to ensure it has no discriminatory impact on any protected characteristic such as race or sex.
Additionally, employers would be required to disclose when an automated decision system is being used, a description of the system, types of data collected by the system, characteristics the system is intended to measure, how the applicant or employee can interpret the system output in plain language, and how the applicant or employee can dispute any decision. Employers would be prohibited from retaliating against those who “sought assistance” or “filed a complaint” against use of an automated decision system.
Enforcement
The Department of Labor (DOL) would enforce the bill through a new administrative process while individuals would have a private right of action. The new bill would establish an expert artificial intelligence (AI) advisory committee for the proposed new DOL division with membership that includes individuals from a range of geographic regions, with employer-side perspectives, and with specific expertise on how AI is used in the workplace.
The bill proposes statutory damages for violations in the range of $5,000 to $20,000 per violation up to $40,000 for willful or repeated violations. Violating the retaliation provisions would increase to $50,000 per violation and $100,000 for willful violations. Additionally, in litigation individuals can recover actual damages, treble damages, and attorneys’ fees.