CA SB 420 (a broad high-risk automated decision system law covering employment, housing, healthcare, education, and lending) and CA SB 947 (an employment-only ADS bill prohibiting protected-status inference and mandating human review for terminations) both cleared Assembly committees this week — if both are enacted, compliance teams will face layered, non-identical obligations in the employment domain, with SB 947's private right of action applying on top of SB 420's AG/CRD-only enforcement.

CA SB 1119 (comprehensive child-safety obligations including annual risk assessments and ephemeral-by-default settings for minors), CA SB 300 (amending enacted SB 243's knowledge trigger to codify a constructive-knowledge standard), and CA SB 867 (banning companion chatbots embedded in toys) all advanced in the California Assembly this week — collectively tightening SB 243's regime across child-safety scope, product-form, and knowledge-standard dimensions simultaneously.

MI HB 5570 passed both chambers and was enrolled on July 3, adding dissemination of intimate deepfakes with aggravating factors (Class F felony, 3-year max) to Michigan's sentencing guidelines, but it creates no compliance obligations for AI companies and expressly does not take effect unless companion bill HB 5569 (which would create the underlying substantive offense) is also enacted.