4 Comments
User's avatar
The Artificial Intelligence's avatar

AB 1883 lines California up with Europe on one point. The EU AI Act has banned emotion recognition in workplaces and schools since February 2025, apart from medical and safety uses.

Vandana Gyanchandani's avatar

The contrast between preserving clinical decision-making and vetoing protection for nurses who challenge unsafe AI is striking. A professional can formally remain the decision-maker while facing strong pressure to accept the recommendation. For oversight to mean much, people need time to review the evidence, authority to reject the output and protection when they raise a safety concern. I’d judge implementation by what happens to the person who says no, not simply by whether a human signs the record.

Jillian Berk's avatar

A human oversight requirement reads as a limit on automation. Operationally it is a relocation of the signature. The system still produces the recommendation. A licensed person is now required to endorse it, and that person carries the liability the vendor would otherwise have had to price. Rules like these make the tools cheaper to sell into regulated settings, not harder. Watch adoption in clinical work over the next two years.

AI × World of Work's avatar

Full of useful information and insights. Really appreciate this!