California has disclosed the use of six high-risk automated decision-making systems after initially reporting none. The systems include recidivism prediction, fraud detection, and exam proctoring.
The report follows a state law requiring annual transparency. The findings highlight ongoing debates over AI safety and oversight.
- The state initially denied using high-risk AI systems, but a new report reveals six in use. - Systems include COMPAS for recidivism scores and a fraud detection tool that paused 600,000 benefits.
- The report excludes generative AI pilots like Poppy and ChatGPT contracts. - Courts in Los Angeles and Riverside counties are testing AI for clerical tasks.
The disclosure comes amid public concern over AI safety and a push for more transparency in government use of technology.