On October 19, 2026, New Jersey Assembly Bill A5607, introduced on October 5, cleared the Assembly Public Safety and Preparedness Committee. It now sits in Appropriations and has not become law. Legal and tech press picked it up this week, with the ABA Journal reporting on it on November 3 and Gadget Review on November 5, because of one provision: it would classify AI chatbots as products under New Jersey's product-liability law. Under Gadget Review's reading of the bill text, a provider would be liable for "any injury to a user from the use of its chatbot," even if it "exercised all reasonable care" in design and distribution, and even if it had no direct contract with the injured user.

Why It Matters

**AI companion apps would be most exposed.** Romantic and sexual companion apps hold the most intimate chat histories of any consumer AI product. Many make money through engagement-driven upsells and, in some cases, ad-tech, and both would come under strict-liability exposure. A strict-liability standard also makes the kind of harm claims that helped push Character.ai into Chapter 11 ([our coverage](/stories/character-ai-chapter-11-disney)) easier to bring: a plaintiff would not have to prove negligence, only that the product caused the injury. **It goes further than other state bills.** Most state companion-bot laws so far require disclosure, crisis protocols and protections for minors ([our twelve-state map](/stories/twelve-state-companion-bot-laws-map)), and New York's new law focuses on minors ([our coverage](/stories/ny-chatbot-minors-safety-law)). A5607 applies product liability to **all users**, adults included. It still has to pass Appropriations and both chambers, and heavy industry lobbying is likely. If it passes even in amended form, expect companion-app operators to geofence New Jersey, add stricter adult-content guardrails, or require arbitration terms that the statute may not allow.

The money. The attorney general could seek civil penalties of $5,000 per violation and court orders to stop or change the challenged conduct. Users harmed by covered privacy violations could sue in Superior Court for the greater of actual damages or $5,000 per violation, plus attorney fees. "When there's $5,000 in statutory damages per violation and you don't have to prove any harm, in addition to the strict liability standard, you're talking hundreds of thousands and perhaps millions of dollars per lawsuit," Scott S. Christie, a McCarter & English partner, told Law.com, as quoted by the ABA Journal. "Because if the chatbot provider is violating the statute as to one user, it's probably violating as to all users." He said the bill would "open the floodgates."

The data rules. According to Gadget Review, A5607 would ban using chat logs for advertising or behavioral profiling, ban selling a user's conversation history, restrict some AI-training uses of chat data, and protect users who refuse consent from retaliation. It would also generally cap retention of chat logs at two years, let users download their logs at any time, and require monthly risk assessments and a full data-security program. Gadget Review notes a tension: the bill limits how conversations can be monetized while also requiring providers to keep them. The Electronic Privacy Information Center supports the bill.

Sources

Update — 2026-11-06

Initial entry — story first created.