A coalition of state attorneys general is probing OpenAI across two distinct but converging fronts: the safety of minors on its platforms and the governance terms of its high-profile recapitalization, according to a Bloomberg report published June 13, 2026.

The coordinated scrutiny has been building for months. In August 2025, Arizona AG Kris Mayes joined a bipartisan group of 44 state attorneys general in formally warning major AI companies over predatory design practices affecting children. That same month, California AG Rob Bonta and Delaware AG Kathy Jennings sent a separate, targeted letter to OpenAI specifically on harm to children, putting the company on notice ahead of any formal enforcement action.

Two Tracks, One Company

The child safety thread and the corporate governance thread have run in parallel, occasionally intersecting.

On governance, OpenAI's transition from a nonprofit-controlled structure to a public-benefit corporation drew direct intervention from multiple state regulators with jurisdiction over charitable assets. Bonta issued a statement on OpenAI's recapitalization plan in October 2025. On the same day, Jennings announced the completion of her office's review, stating that the process had secured structural reforms and concrete commitments on safety — Delaware being the state of OpenAI's nonprofit incorporation and therefore carrying unusual legal weight in that review.

The child safety track moved toward something more collaborative in late 2025. North Carolina AGs Jeff Jackson and Derek Brown launched a nationwide bipartisan AI task force in November 2025, with OpenAI and Microsoft participating directly — a less adversarial posture than a formal probe, though not mutually exclusive with one.

That collaborative framing produced a concrete output. In April 2026, OpenAI published a Child Safety Blueprint co-developed with Jackson and Brown, who serve as co-chairs of the Attorney General Alliance Artificial Intelligence Task Force. The document lays out OpenAI's stated framework for protecting minors across its products and APIs. Publishing it with AG co-authorship is a meaningful credential — it extends political cover to OpenAI while giving participating AGs a tangible deliverable to point to.

What the Probe Is Actually Testing

The mechanics of state AG authority here are worth unpacking. Attorneys general can act under consumer protection statutes, charitable trust laws (relevant where nonprofit assets are involved), and, increasingly, state-level children's online privacy frameworks modeled on COPPA but with broader enforcement teeth. The 44-state coalition letter from August 2025 likely invoked several of these simultaneously — a common tactic to maximize jurisdictional surface area.

OpenAI's own internal posture on accountability is documented in its Raising Concerns Policy, dated January 12, 2026, which explicitly prohibits retaliation against employees who surface concerns or cooperate with investigations. That policy now sits in the background of an active multi-state probe — which means any allegation of whistleblower pressure would carry additional legal exposure.

The broader context here is that this is not a story about one regulator taking a unilateral swing at a high-profile target. Forty-four states coordinating a warning letter, two AGs reviewing a multi-billion-dollar recapitalization, and a bipartisan task force producing a co-branded safety document represent a regulatory apparatus that is simultaneously adversarial and transactional — probing for leverage while keeping a channel open for compliance.

OpenAI, for its part, has chosen engagement over stonewalling at almost every juncture: co-authoring the child safety blueprint with the same AGs whose offices are scrutinizing it, allowing Delaware to review its restructuring in exchange for sign-off, and participating in the NC-led task force. Whether that posture reflects genuine alignment with regulators or a calculated effort to shape the terms of oversight before federal frameworks arrive is a fair question — one whose answer will likely depend on what the current probe surfaces.

The probe is ongoing. No enforcement actions or formal findings have been publicly announced as of June 13, 2026.