• Florida Attorney General seeks to bar OpenAI from developing new AI models without outside oversight and to prohibit ChatGPT access for minors.
  • A separate proposed nationwide class action alleges OpenAI collected children's sensitive data without parental consent and seeks a proportionate equity interest in the company.
  • OpenAI disputes liability, points to newly paused model training and improved safeguards, and continues to face a multi-company FTC inquiry and international litigation.

OpenAI is confronting a widening legal assault over the safety of its ChatGPT chatbot for children, with a Florida enforcement action and a newly filed nationwide class action intensifying scrutiny on the company's data practices and product design. The cases, while distinct, both target the company's handling of minors and could reshape how generative AI services are regulated.

Florida Seeks Sweeping Restrictions

On September 28, Florida Attorney General James Uthmeier asked a court to prevent OpenAI from developing new AI models without outside oversight, prohibit ChatGPT access for minors, and stop the company from giving its chatbot "human attributes." The request for an injunction stems from a June 1 lawsuit alleging misleading safety claims, addictive use, and assistance related to self-harm and violence. If granted, the restrictions would mark one of the most aggressive state interventions into AI development to date, potentially forcing changes to core aspects of ChatGPT's design and availability.

OpenAI has pushed back, saying it had paused training its most capable models pending additional safeguards. On October 14, CFO Sarah Friar reiterated that a next-generation model was being held back for additional safety and alignment work. The company has also argued that information cited in individual harm cases is broadly available online and that it continues improving safety tools.

The legal distinction matters: Florida's action is a request for an injunction, not a confirmed court-ordered ban. The underlying lawsuit alleges harm, but no final determination of liability has been made.

New Class Action Targets Children's Data

A separate proposed nationwide class action, filed October 13 in federal court in San Francisco, alleges that OpenAI collected minors' sensitive information without required parental consent and used it to train models and build profiles. The complaint seeks an unusual remedy: a proportionate equity interest in OpenAI, or equivalent cash, for children whose data allegedly contributed to its models.

The privacy case challenges whether parental-consent protections work in practice. OpenAI's terms prohibit use by children under 13 and require parental consent for users under 18; the complaint alleges those protections were insufficient. The class action has not yet been certified, and the allegations remain unproven.

Company Under Pressure on Multiple Fronts

The legal developments come as OpenAI navigates a complex financial and operational landscape. The company reported more than 900 million weekly ChatGPT users and over 50 million subscribers as of March 31, when it announced $122 billion in committed capital at an $852 billion post-money valuation. Annualized revenue was reported near $50 billion at the end of September, according to financial documents cited by the Financial Times, though that figure excluded revenue from cloud partners, according to a person familiar with the matter. OpenAI has not yet turned a profit.

The company is targeting roughly $600 billion in total compute spending through 2030, underscoring its dependence on continued revenue growth and financing. Infrastructure projects in Ohio, Texas and Georgia connect its expansion to local construction and investment. A substantial development restriction from the Florida case could complicate those plans, though no project cancellation attributable to the litigation has been established.

OpenAI completed its recapitalization in October 2025, with the nonprofit OpenAI Foundation retaining control of the commercial business, OpenAI Group PBC. That structure pairs commercial fundraising with a stated public-benefit mission but does not resolve the safety allegations.

Leadership changes have also drawn attention. On October 14, infrastructure lead Sachin Katti announced he would leave his operational role on November 15 while remaining an adviser; Uday Ruddarraju, the company's CTO of compute, will take over the compute organization. CNBC reports that product and business chief Fidji Simo moved toward an advisory role in July for health reasons, and that Brad Lightcap and revenue chief Denise Dresser departed in August. The reporting does not establish that these departures resulted from child-safety litigation.

Broader Scrutiny and Industry Implications

The Federal Trade Commission is conducting a separate fact-finding study, having issued Section 6(b) information requests to seven chatbot companies, including OpenAI, Alphabet (GOOGL), Character Technologies, Instagram, Meta (META), Snap (SNAP) and xAI. The inquiry examines safety testing, age restrictions, monetization and compliance with the Children's Online Privacy Protection Act rule. The FTC explicitly described this as a study, not a specific enforcement proceeding.

Internationally, British Columbia sued OpenAI over its alleged failure to notify authorities before the February Tumbler Ridge mass shooting, and families of victims have brought separate cases. This extends the debate across borders, particularly around when a chatbot provider should alert local law enforcement.

The societal stakes are high. The cases raise concerns about self-harm guidance, emotional dependence, addictive use and disclosure of sensitive personal information. For parents, the privacy case challenges whether parental-consent protections work in practice. For schools and educators, Florida's requested prohibition on minors' access would affect educational use as well as harmful use, creating a trade-off between access and protection.

OpenAI says governments should establish pragmatic standards applying to the entire AI industry rather than one company. The FTC identifies the capacity of companion chatbots to mimic human emotions and intentions as a reason to examine their impact on young users.

What to Watch

The consequential developments in the near term are a ruling on Florida's requested injunction, whether the October 13 privacy suit receives class certification, and whether discovery reveals evidence about safety testing, age controls and internal knowledge of risks. OpenAI's model-safety pause is already an operational development, but it should not be described as a court-imposed restriction or attributed solely to Florida's lawsuit.

A central legal issue is whether courts treat these cases primarily as disputes over information, defective product design, deceptive marketing or children's privacy. That classification could shape both liability and remedies across the industry. The legal commentary retrieved highlights these competing theories but does not support a confident prediction of the outcome.

OpenAI did not respond to a request for comment on the pending litigation. A spokesperson previously said the company is committed to improving safety tools and engaging with regulators.

Correction: An earlier version of this article misstated the filing date of the proposed class action. It was filed October 13, not October 14.