Key Takeaways
- A coalition of state attorneys general, led by New York and Colorado, has subpoenaed OpenAI for internal documents concerning user data, minor safety, and advertising practices.
- OpenAI says it takes the concerns seriously, points to new safeguards in ChatGPT (e.g., parental controls), and pledges to improve.
- The investigation adds to a wave of state‑level AI scrutiny, including legislation, executive orders, and lawsuits targeting child safety, copyright infringement, and labor impacts.
- Federal actions under the Trump administration have also tightened oversight, barring foreign access to certain AI models and urging voluntary pre‑release government review.
- Precedents such as New Mexico’s $375 million verdict against Meta show that state investigations can lead to costly, prolonged litigation.
- Other states are pursuing similar actions against AI companions and deep‑fake technologies, signaling a broad, multi‑front effort to regulate emerging AI risks.
State Attorneys General Launch Investigation into OpenAI
On Friday, attorneys general from New York, Colorado, and several other states served subpoenas on OpenAI demanding internal records. The requests focus on how the company handles user data, protects minors from harmful content, and conducts its advertising activities. Two sources familiar with the probe confirmed the involvement of multiple states, speaking on condition of anonymity due to the ongoing legal matter. The subpoenas mark a significant escalation in state‑level scrutiny of the AI startup’s operations.
OpenAI’s Response and Commitment to Safety
OpenAI issued a brief statement acknowledging the seriousness of the concerns raised by the state attorneys general. The company said it intends to engage constructively with their offices and emphasized that the latest version of its model, ChatGPT, now includes additional safeguards such as parental controls designed to protect children. OpenAI added that while it cannot erase past harms, it is committed to learning, improving, and getting its safety measures right moving forward.
Broader Context: Rising AI Concerns and Harm
The investigation comes amid growing public alarm over AI‑related harms. Reports have linked increased instances of self‑harm among children to interactions with AI chatbots, while AI‑generated scams have proliferated. Critics also warn about the technology’s potential to displace workers, the massive energy consumption of data centers that power AI systems, and the erosion of copyright protections as models train on vast troves of online content. These anxieties have fueled both federal and state calls for tighter oversight.
Federal Actions Under the Trump Administration
The Trump administration has taken steps to address national security risks associated with AI. On Friday, it barred the AI startup Anthropic from granting foreign nationals access to its new Mythos and Fable 5 systems. Earlier in the month, President Trump signed an executive order urging technology firms to voluntarily submit new AI models for government review before public release. Although the approach remains largely advisory, it signals a shift toward greater federal involvement in AI governance.
State‑Level Legislative Surge on AI Guardrails
State legislatures have been exceptionally active this year, introducing dozens of bills aimed at placing guardrails around AI development and deployment. More than 100 state laws now exist that prohibit chatbots for young users, mandate security‑risk testing of AI systems, and protect copyrighted material from being scraped for model training. This flurry of legislation reflects a bipartisan effort to pre‑emptively address perceived dangers before they become widespread.
California’s Proactive Labor Policy Executive Order
Governor Gavin Newsom issued an executive order last month directing state agencies to collaborate with academics, labor groups, and AI industry representatives to study how to subsidize companies that retain human workers instead of replacing them with automation. The order seeks to mitigate potential mass job displacement caused by AI and to explore policy tools that could support workforce transition in an AI‑driven economy.
Illustrative Precedents: State Lawsuits Against Tech Firms
State investigations can evolve into costly, drawn‑out litigation. In 2023, New Mexico’s attorney general, Raúl Torrez, sued Meta over child‑safety failures on its platforms, ultimately securing a $375 million jury award. Florida became the first state to sue OpenAI this month, alleging that its ChatGPT chatbot endangered children and that the company failed to disclose those risks publicly. Florida’s attorney general has also opened a criminal investigation into OpenAI after reviewing messages between the chatbot and the individual accused of a deadly shooting at Florida State University last year.
Specific Allegations in Florida’s Lawsuit Against OpenAI
Florida’s lawsuit contends that OpenAI’s ChatGPT poses a direct risk to minors by facilitating harmful interactions and that the firm neglected to warn users about these dangers. The state argues that the lack of adequate disclosures violates consumer‑protection statutes and that the company’s design choices have enabled unsafe use by children. The complaint seeks injunctive relief, damages, and mandatory safety reforms, highlighting how state attorneys general are using existing legal tools to challenge AI products.
Other State Actions: AI Companions and Deepfake Concerns
Beyond chatbots, states are targeting other AI applications. California’s attorney general, Rob Bonta, announced an investigation into x.AI—a SpaceX unit—for producing non‑consensual sexualized deep‑fake images of real women and minors. Kentucky has sued Character.AI, a maker of AI companions, alleging that its products expose children to dangerous or abusive uses in violation of state consumer‑protection laws. These cases illustrate a widening net of state enforcement that spans generative media, companionship bots, and deep‑fake technologies.
Conclusion: Implications for AI Regulation and Industry Practices
The cascade of state investigations, lawsuits, and legislative proposals signals a pivotal moment for AI governance. Companies like OpenAI may face prolonged legal battles, substantial financial penalties, and compelled changes to their product designs and data practices. As states experiment with varied approaches—from child‑safety protections to labor‑preservation subsidies—the tech industry will need to adapt quickly to a fragmented but increasingly assertive regulatory landscape. The outcomes of these efforts could shape national standards and determine how swiftly AI innovations can be deployed while safeguarding public welfare.