Politics & Government

LA County Sets New Rules For AI, Bans Fully Automated High-Stakes Decisions

The policy sets countywide rules for privacy, transparency, human oversight and the use of sensitive data.

The policy says AI should be used to enhance jobs and provide training and retraining opportunities for workers, rather than replace county employees or contractors.
The policy says AI should be used to enhance jobs and provide training and retraining opportunities for workers, rather than replace county employees or contractors. (Chris Lindahl/Patch)

LOS ANGELES, CA — The Los Angeles County Board of Supervisors approved a policy Tuesday governing the safe, ethical and transparent use of artificial intelligence technologies across county government.

The policy establishes standards, responsibilities and governance requirements for AI systems developed by the county, procured from outside vendors or accessed through third-party platforms.

"AI technologies are increasingly used across county departments to enhance service delivery, improve operational efficiency, and support data driven decision making," according to a board letter from the county's chief executive officer. "As these technologies advance, it is critical for the county to maintain clear governance structures, safeguard sensitive information, and ensure that AI is deployed responsibly and ethically."

Find out what's happening in Los Angelesfor free with the latest updates from Patch.

The policy requires county departments to follow principles involving human-centered design, security and safety, privacy, transparency, equity, accountability, effectiveness and workforce empowerment.

"The AI Policy establishes consistent requirements to ensure that any AI technologies developed, procured, or used by county departments comply with applicable laws, protect county information assets, maintain public trust, and mitigate the risk of bias or misuse," the board letter states.

Find out what's happening in Los Angelesfor free with the latest updates from Patch.

Under the policy, AI systems must protect personally identifiable information, protected health information, criminal justice information and other sensitive data from unauthorized access, disclosure and manipulation.

The policy also requires the county to disclose the purpose and use of its AI systems to the public and identify work products generated by AI.

It prohibits fully automated high-impact decisions without meaningful human review, covert tracking or social scoring of people, and the cognitive or behavioral manipulation of people or vulnerable groups.

County employees are also barred from uploading nonpublic or regulated county information to unapproved AI platforms, using such platforms to make official decisions or relying on AI-generated work as an authoritative source without independent human verification.

The policy says AI should be used to enhance jobs and provide training and retraining opportunities for workers, rather than replace county employees or contractors.

It assigns oversight responsibilities to the county's chief information officer, chief information security officer, chief privacy officer and department-level information technology managers.

The policy allows no exceptions, according to the county.

County Counsel reviewed the proposal, and the Audit Committee approved it Sept. 16. "There will be no fiscal impact to the county on the proposed AI Policy," according to the board letter.

City News Service