Newsom Signs Social Media Liability and Teen Feed Limits and Pauses Chatbot Toys; New York Ends File-and-Use Auto Rate Increases

The bottom line today

The ceremony was held at a children’s museum near San Francisco [3]. What Gov. Newsom signed there Thursday is a package of 13 bills that, for the companies on the receiving end, amounts to a new liability regime for social media and a new compliance regime for chatbots [1, 3]. The goal of protecting children online is one nobody in business argues with. The tools are the question: a damages provision built on negligence, a feature ban whose definitions its critics call too broad, and a four-year pause on a product category [3, 4]. In Albany, the Department of Financial Services published the rule that ends file-and-use rate increases for personal auto insurance [6, 7]. Both states are adding cost. Only one of them paired it with relief.

CA California AB 2: negligence damages for harm to children

What it does. AB 2, by Assemblymember Josh Lowenthal, is listed by the governor’s office as civil penalties for injuries to children [1]. The Associated Press reported that it lets large social media companies be penalized up to $1 million per child if they are found negligent in harming children through their platforms [3].

Where it stands. Signed September 10 [1].

Pratt’s read. This is the bill in the package with the widest cost, because it works through the courts rather than a regulator. Lowenthal’s case, as he put it to the AP, is that “the pace of technology is moving faster than government can put guardrails in” [3]. That is a reasonable description of the problem. A per-child penalty attached to a negligence finding is a reasonable description of a litigation magnet. The size of the exposure depends on how a court applies negligence to a feed design, and nothing in the record this release draws on estimates it. Expect the plaintiffs’ bar to read the statute before the platforms’ compliance teams do.

CA California AB 1709: addictive features off for users under 16

What it does. AB 1709, also by Lowenthal, restricts addictive design features for users under 16 [2]. Under the law, platforms must either stop offering tools such as infinite scroll and autoplay to those users or keep them off the platform entirely, CalMatters reports [4]. The governor’s office describes the ban as reaching features such as autoplay and algorithmic feeds based on user history and profile, and AB 1709’s title adds an e-Safety Advisory Commission [1].

Where it stands. Law as of this week; the governor’s office lists it among the 13 bills in the package [1].

Pratt’s read. Meta objected that restricting feeds hinders what makes its apps useful. “Personalization is also how we deliver age-appropriate content for teens that is relevant to them — all with the proper guardrails in place,” spokesperson Jim Cullinan said [3]. The Electronic Frontier Foundation, not a business lobby, urged a veto and called the bill “well-intentioned, but deeply flawed,” arguing that its definition of addictive features is broad enough to work as a full ban for users under 16 [4]. When the civil liberties groups and the platforms reach the same conclusion, the drafting deserves a second look. For any company with a feed, the practical cost is building and maintaining a separate under-16 product, plus knowing who is under 16. Another bill in the package, AB 1856, deals with age verification signals for software applications [1], and CalMatters reports that one of the signed laws expands and revises age-checking requirements [4].

CA California SB 1119: audits for companion chatbots

What it does. SB 1119, by Sen. Steve Padilla with Assemblymembers Buffy Wicks and Rebecca Bauer-Kahan, is the bill supporters call Adam’s Law [1, 2]. The governor’s office says it requires crisis protocols for suicidal ideation, parental controls and notice when a child disables safety settings, and that it is the first law in the country to require companies to conduct independent child safety audits and annual risk assessments [1]. CalMatters reports that it imposes time limits for minors, requires the operator to alert parents when a chatbot detects a threat of self-harm, and leaves companies that fail to act open to legal liability [4]. The AP describes the package as requiring chatbot operators to perform risk assessments before rolling products out [3].

Where it stands. Signed alongside the others, with co-authors from both houses [1, 2].

Pratt’s read. Credit where due: this is the bill that drew industry support. OpenAI said the law “pairs strong protections with continued access to useful AI tools,” according to CalMatters [4]. Safety plans and crisis protocols are things a responsible operator should already have. The cost sits in the audit requirement, which turns a one-time design decision into an annual outside engagement. As this release noted on September 8, California is also writing rules for AI used to discipline and fire workers (earlier release); chatbot operators now have their own audit calendar to add.

CA California SB 867: a pause on chatbot toys

What it does. SB 867, by Padilla, places a moratorium on AI-powered chatbot toys for children [2]. CalMatters describes it as a four-year moratorium covering kids under 16 [4].

Where it stands. Signed at the same ceremony [1, 3].

Pratt’s read. A moratorium is the bluntest tool on the list. It pauses a product category in California before the state has set a safety standard for it, and toy makers and retailers lose that market for the length of the pause. The same package shows the alternative: for companion chatbots used by children, SB 1119 sets protocols and audits rather than a ban [1]. A standard with a compliance path costs less than a four-year wait.

NY New York 11 NYCRR 163: every personal auto rate increase needs approval

What it does. The Department of Financial Services published a proposed second amendment to 11 NYCRR 163, the flexible rating rule, on September 9 [6, 7]. It implements Part II of Chapter 55 of the Laws of 2026, signed May 27, which amended Insurance Law section 2350 so that an insurer may no longer raise overall average rates by up to 5% without the superintendent’s prior approval [6]. Decreases of up to 5% can still be filed and used [6]. Until now, insurers could take up to two increases a year without prior approval if the cumulative effect stayed within 5% [7]. Insurers that cut rates because of the budget’s reforms must tell policyholders so at renewal, 30 to 60 days before the policy period ends [6]. The rule covers nonbusiness policies, defined as vehicles used predominantly for nonbusiness purposes with a natural person as the named insured, and for-hire vehicle policies [6]. The amended text still contemplates file-and-use increases on for-hire policies, capped at 30% a year per policy [6].

Where it stands. Proposed; the comment period runs to November 9, and the law and the rule take effect November 27 [7, 8].

Pratt’s read. The relief came first, and it is real. The same budget narrowed the “serious injury” threshold for pain-and-suffering suits, limited payouts to drivers found more than 50% at fault, and widened criminal liability for organizers of staged accidents [8, 9]. Those are cost cuts, and DFS has told insurers to build the expected savings into pending and future filings [8]. Supporters’ case for prior approval is accountability. “New Yorkers deserve a transparent, fair and accountable auto insurance market that puts customers first,” Gov. Hochul said [7]. Assemblymember David Weprin called the requirement “an important consumer protection” [7]. Lawmakers, not the governor, insisted on it, and the insurance industry, which backed the reform package, may not welcome it, Insurance Journal reported in May [9].

The business read is narrower than the headlines. Fleet and commercial auto policies are outside the rule; a contractor’s trucks are not covered [6]. The owner who drives a personal car to job sites is. For insurers, the change trades a 5% band they could move within for a queue. A filing that waits on approval cannot respond quickly to a bad quarter, while DFS has already told insurers to price in savings the reforms are only expected to produce [8].

On the calendar

Comments on New York’s proposed amendment to 11 NYCRR 163 are due November 9 [8]. Prior approval becomes mandatory on November 27 [6, 7]. In California, the next step for AB 2, AB 1709, SB 1119 and SB 867 is implementation, starting with the e-Safety Advisory Commission AB 1709 creates [1].