A coalition of state attorneys general is investigating OpenAI on two related fronts: whether its platforms harm minors and whether its corporate restructuring followed proper governance rules, according to reporting from June 2026.

The scrutiny has been building steadily. In August 2025, Arizona AG Kris Mayes joined a bipartisan group of 44 state attorneys general in formally warning major AI companies that they were using predatory design practices—features intentionally engineered to be addictive or manipulative—that affected children. The same month, California AG Rob Bonta and Delaware AG Kathy Jennings sent a separate letter directly to OpenAI warning of potential harm to minors, signaling that enforcement action could follow if the company did not address the issue.

Two Separate but Connected Investigations

The child safety investigation and the corporate governance investigation have run side-by-side, sometimes overlapping.

On the governance side: OpenAI was originally structured as a nonprofit organization, which meant it held charitable assets under the oversight of state regulators. When the company began moving toward a different structure as a public-benefit corporation, this triggered review by attorneys general in states with jurisdiction over those assets. In October 2025, California's Bonta issued a statement on OpenAI's restructuring plan. The same day, Delaware's Jennings announced she had completed her office's review, saying it had secured structural reforms and safety commitments from OpenAI. Delaware is significant here because OpenAI's nonprofit was incorporated there, giving the state unusual legal authority in the matter.

The child safety investigation moved in a less adversarial direction in late 2025. North Carolina AGs Jeff Jackson and Derek Brown launched a bipartisan national task force on AI in November 2025, with OpenAI and Microsoft participating directly. This was more collaborative than a formal probe, though the two can coexist.

That collaborative approach produced a tangible result. In April 2026, OpenAI released a Child Safety Blueprint that was co-developed with Jackson and Brown, who co-chair the Attorney General Alliance Artificial Intelligence Task Force. The document outlines how OpenAI says it will protect minors across its products and services. The fact that state attorneys general co-authored this document matters—it gives OpenAI political credibility while giving the participating AGs a concrete result to show their constituents.

How State Attorneys General Wield Power

State attorneys general have legal tools available for cases like this. They can pursue cases under consumer protection laws, charitable trust laws (relevant when nonprofit assets are at stake), and increasingly under state-level laws designed to protect children online. These laws are modeled on a federal rule called COPPA, but with broader enforcement power. The 44-state warning letter likely invoked several of these legal frameworks at once—a standard approach to maximize the jurisdictional reach.

OpenAI has also published an internal policy, dated January 2026, that explicitly protects employees who report concerns or cooperate with investigations from retaliation. That policy now sits in the background of this multi-state investigation, meaning any case of an employee being pressured would carry additional legal risk for the company.

The broader pattern here is instructive. This is not simply one regulator going after a high-profile company. Forty-four states sending a coordinated warning letter, two attorneys general reviewing a multibillion-dollar restructuring, and a bipartisan task force producing a jointly-branded document on safety—all of this shows a regulatory system that mixes skepticism with negotiation. The attorneys general are investigating OpenAI while also keeping channels open for the company to comply and address concerns.

OpenAI's response strategy has been to engage rather than resist. The company co-authored the child safety blueprint alongside the same regulators investigating it. It allowed Delaware to review its restructuring in exchange for approval. It participated in the North Carolina-led task force. Whether this reflects genuine agreement with regulators or a deliberate effort to shape oversight before federal rules arrive is a fair question. The answer will likely become clearer as the current investigations proceed.

As of mid-June 2026, the investigations remain open. No enforcement actions or formal conclusions have been announced.