OpenAI Faces Multistate Scrutiny as State Attorneys General Probe AI Safety Practices

OpenAI is facing a fresh wave of legal and regulatory scrutiny after a coalition of state attorneys general reportedly began investigating the company’s consumer AI practices. The probe, first reported by The Wall Street Journal and summarized by TechCrunch, adds a new state-level front to questions already surrounding ChatGPT’s safety systems, data practices, product design, and impact on vulnerable users.

According to TechCrunch, OpenAI was served with a subpoena from New York’s attorney general on Friday. The document request reportedly covered a broad set of topics: advertising, user engagement and retention, model sycophancy, consumer data, health data, and how the company treats minors and seniors. The states involved have not been publicly identified, and OpenAI has not disclosed detailed information about the request.

What investigators appear to be examining

The reported subpoena points to a wider concern that modern AI chatbots are no longer simply productivity tools. They are consumer products used for work, education, emotional support, health questions, and personal decision-making. That makes product design choices—how a chatbot responds, how it retains users, how it handles sensitive disclosures, and how it protects children—central to state consumer-protection enforcement.

One notable area is “model sycophancy,” a term used to describe AI systems that may overly agree with users, reinforce mistaken beliefs, or validate harmful thinking instead of challenging it safely. Regulators appear increasingly interested in whether chatbot behavior can create risk when users are distressed, young, elderly, or seeking health-related advice.

Data handling is another likely focus. Chatbot interactions can include highly personal information, including mental-health concerns, medical symptoms, family details, school issues, or financial questions. If state officials are requesting documents about consumer data and health data, the investigation may examine both OpenAI’s disclosures to users and the internal controls used to protect sensitive information.

OpenAI says it will engage with the states

OpenAI told TechCrunch that it takes the concerns raised by state attorneys general seriously and intends to engage constructively. The company also said ChatGPT now includes more protective experiences for minors and people experiencing difficult situations, including safeguards that direct users toward real-world resources and trusted human contacts.

The company has emphasized child-safety work such as age prediction, parental tools, and restrictions on advertising that targets kids. Those steps may now be evaluated not only as product improvements, but as evidence in a broader debate over whether the industry’s safeguards have kept pace with adoption.

The investigation follows Florida’s lawsuit

The multistate scrutiny comes shortly after Florida Attorney General James Uthmeier announced a lawsuit against OpenAI and CEO Sam Altman. Florida’s official release alleges that OpenAI knowingly released and aggressively marketed ChatGPT while concealing serious risks, suppressing safety warnings, and deceiving Floridians about the product’s dangers. OpenAI has disputed claims that it ignores safety and has said it continuously strengthens safeguards to detect harmful intent, limit misuse, and respond when risks arise.

The Associated Press reported that Florida’s complaint references cases in which alleged gunmen were said to have asked ChatGPT questions while planning crimes. OpenAI said its models encouraged the individuals to seek real-world support and that the company cooperated with law enforcement. Those facts will matter because regulators are likely to distinguish between misuse of a general-purpose tool and allegations that product design, warnings, or safety systems were inadequate.

Why the timing matters

The state investigation also arrives as OpenAI is preparing for a potentially historic public-market debut. CNBC reported that the company has confidentially filed for an IPO, while continuing to operate in an expensive and highly competitive AI infrastructure race. That timing raises the stakes: prospective investors will be weighing not only growth, revenue, and model performance, but also litigation exposure, compliance costs, and reputational risk.

For the wider AI sector, this inquiry may signal a shift from broad policy debates to document-heavy enforcement. State attorneys general often use consumer-protection laws to scrutinize marketing claims, privacy practices, youth safety, addiction-like product design, and unfair or deceptive business practices. If the coalition expands or publishes findings, other AI companies could face similar questions about retention mechanics, health-related interactions, and protections for children.

What to watch next

The immediate question is which states are involved and what the subpoena ultimately reveals. Key indicators will include whether investigators focus mainly on disclosures and advertising, or whether they examine model training, response behavior, safety escalation, and internal risk assessments. Another important sign will be whether the states coordinate with existing lawsuits or pursue separate settlement demands, product changes, or penalties.

OpenAI’s response will likely emphasize updated safeguards, parental controls, age prediction, and cooperation with regulators. But the larger issue is now industry-wide: as AI assistants become everyday interfaces for search, advice, education, and personal support, regulators are asking whether companies can prove that rapid deployment has not outpaced consumer safety.