Using AI to make termination decisions in the workplace may be limited next year
The measure — the No Robo Bosses Act of 2026 — awaits Gov. Newsom’s signature or veto
Federal and state lawmakers are struggling with how to constrain the evolving hazards of artificial intelligence that are moving faster than the speed of law, a phrase I once applied to the coronavirus.
Sitting on Gov. Gavin Newsom’s desk awaiting signature or veto as of this writing is Senate Bill 947, or the No Robo Bosses Act of 2026. If signed, the measure would restrict employers’ use of automated decision systems (ADS) in making termination or disciplinary decision starting July 1, 2027.
ADS defined, restricted
The measure defines automated decision system as “any computational process derived from machine learning, statistical modeling, data analytics, or artificial intelligence that issues simplified output, including a score, classification, or recommendation, that is used to assist or replace human discretionary decisionmaking and materially impacts natural persons.”
Virtually all public and private employers would be prohibited from relying exclusively on ADS to make disciplinary or termination decisions.
An employer that relies “primarily on ADS output” to make such a decision would have to “direct a human to corroborate the decision using data that was collected or used toproduce the ADS output or other relevant corroborating or supporting information.” “ADS output” is defined as “information, data, assumptions, predictions, scoring, recommendations, decisions, or conclusions generated by an ADS.”
Inputs the ADS may be asked to review include employee evaluations, personnel files, employee work product, peer reviews, and witness interviews, including with the employer’s customers or clients.
If the employer cannot corroborate the ADS’ bottom line, or the human reviewer concludes the ADS output is “inaccurate, incomplete, or misleading” – to what degree is unclear —the employer could not use the ADS output to make the disciplinary decision. The employer then presumably would have to rely on conclusions and recommendations of humans alone to make the decision.
Justification, objections
According to the author of the measure, the requirement of “human oversight and independent verification” of disciplinary and termination decisions will reduce the risk of subjecting employees to dangerous working conditions and unjust termination that may result from the use of ADS alone.
The California Chamber of Commerce and other opponents question whether the measure may thwart the development of ADS technology that can combat human bias in workplace decision-making. Put differently, the human guardrail the measure requires may compromise the neutrality and efficiency for which the best AI is designed.
Employee rights
An employer that has primarily used an ADS to make a disciplinary or termination decision would be required to provide the employee, upon request, “a meaningful, objective description of the employee’s own data used by ADS,” that is, the ADS inputs. Nothing in the measure prohibits using AI to generate that description. The employer also would have to provide any affected employee with a separate written “post-use notice” when the employer informs the employee of the decision. The notice would have to be easily understandable and easily accessible.
The notice would have to contain:
- A statement that the employer primarily relied upon an ADS in making the decision.A statement that a human reviewed the decision and corroborated the ADS output.
- Contact information for a human the employee may reach for more information about the decision and the employee’s right to access a description of the employee’s own data ADS used to make the decision.
- A statement that the employee would not be subject to retaliation for exercising rights under this law.
Click here to read the full article written by SCMV Shareholder Dan Eaton and published in The San Diego Union-Tribune.