OpenAI faces 42-state probe over ChatGPT safety as IPO nears

The multistate investigation into OpenAI marks a new phase in how frontier AI labs are policed: not just by specialized tech regulators, but by the broad coalition of US state attorneys general that usually go after Big Tobacco or Big Pharma. The timing is brutal for OpenAI—just days after filing confidential IPO paperwork—because anything uncovered in subpoenas will need to be reflected in its S‑1, directly affecting valuation, risk factors, and investor appetite. In effect, safety incidents that were previously framed as PR or ethics issues are being reframed as potential consumer-protection violations.

For the broader race to AGI, this is a clear signal that “move fast and break things” is no longer an acceptable operating model at scale. If 42 attorneys general are now combing through data retention policies, youth protections, and so‑called “sycophantic” model behavior, other labs will assume they’re next. That will push serious players to harden red-teaming, logging, incident response and guardrails before launching new models—especially those approaching AGI‑like autonomy. At the same time, the probe could entrench incumbents with the compliance budgets to survive this level of scrutiny, making it harder for smaller labs to compete at the frontier.