Committee OKs online-safety bill amid questions on AI chatbots, employment screening and regulatory sandboxes

The General Law Committee voted to send SB5, a multi-section online-safety bill, to the floor after a broad discussion about how to regulate AI-driven services, companion chatbots and automated decision-making systems.

Sponsor described the bill as the caucus’s online-safety priority: it would require clearer disclosures for subscriptions, add whistleblower protections tied to definitions of "catastrophic risk," direct state agencies to develop a regulatory sandbox plan, require safeguards for companion chatbots that might encourage self-harm or violence, set disclosure and correction rights for some automated employment decisions, and require digital watermarking for AI-generated images and media.

Committee members welcomed some provisions (digital watermarking, training and workforce programs), asked about who would run the proposed regulatory sandbox and whether existing state agencies have capacity, and pressed on liability when private employers deploy third-party hiring tools. Representative Vigliano asked where liability would fall if an employer used a third-party provider; the sponsor answered that liability questions remain unsettled and that the bill focuses on disclosure rather than imposing new penalties. Representative Turko supported the measure to get it out for further study but reserved the ability to press for more research on the effects of potency/cap changes in other contexts (commenting across concurrent AI and cannabis items).

Why it matters: the bill is intended to protect children from harmful chatbot interactions, give consumers clearer rights regarding automated employment decisions, and increase transparency around AI deployment while creating pathways for innovation (a regulatory sandbox). Committees will continue to refine liability, agency capacity and disclosure details as the bill moves to the floor.