California is moving forward with a broad online-safety push as Governor Newsom signs AB 1709, a package of 13 laws targeting minors’ exposure to harmful tech, including a four-year ban on AI companion toys and parental controls on AI chatbots.

The package also bans behaviorally addictive features for under-16s and imposes a regime where large platforms can face up to a $1 million per-child penalty for negligence harming minors.

An accompanying report questions the feasibility and cost of verification, but provides concrete cost examples tied to a recent probe.

Industry voices are split: OpenAI and Anthropic back safeguards and third-party audits, while others warn about the complexity of audits versus evaluations.

Observers frame AI policy as an ongoing effort among policymakers and industry to stay ahead of regulation, invoking thinkers like Friedman and Klein to contextualize the climate.

A central risk is how rigorously the Commission writes standards; outcomes hinge on the quality of criteria and enforcement, not just technology limits.

Experts warn the field relies on unproven tools and that human time could still be overwhelmed by AI outputs, underscoring verification challenges.

Analysts stress that enforcement resources and regulator capacity will determine whether the measures translate into real change.

Newsom and proponents call for cautious progress on advanced AI development amid safety concerns, while some push for slower development and greater transparency; the administration prioritizes U.S. leadership on AI.

OpenAI spends roughly a fifth of its compute on safety monitoring, highlighting the ongoing cost and resource questions of verification for AI systems.

California’s package aims to spur global platform-design regulation, with similar moves in Canada and Europe, though concrete enforcement details remain vague.

An e-Safety Advisory Commission will study and propose legislative changes, though critics argue it lacks sufficient First Amendment expertise.