Florida AG asks judge to bar OpenAI from new AI models
The legal battle over artificial intelligence is heating up, as Florida’s Attorney General has taken a significant step, filing a motion for a temporary injunction against OpenAI and CEO Sam Altman. This move transforms an existing lawsuit into a direct legal challenge concerning the fundamental safety and ethical development of the technology.
The motion isn’t just a formality; it sets forth a series of aggressive demands aimed at controlling how OpenAI operates, particularly regarding safety and data privacy. The request seeks court-ordered stop orders concerning several critical areas: requiring independent third-party guardrails before developing new AI models, banning the provision of ChatGPT to minors in Florida, stopping the collection of data from Florida children under thirteen without explicit consent and robust security procedures, and halting any claims that the service is safe, accurate, or reliable.
Beyond immediate protections for Florida residents, the motion also targets how the service interacts with users, seeking to stop the system from falsely attributing human characteristics to its services and curbing engagement tactics like “conversation prolongation.” This broad scope reflects a deep concern that the unchecked evolution of AI requires immediate, external regulation.
This action is interwoven with Florida’s original lawsuit against OpenAI and Altman, which was filed months earlier and accuses the company of deceptive trade practices, negligence, and gross negligence. The temporary injunction serves as a powerful mechanism to enforce these claims while addressing urgent safety concerns that have recently surfaced in the AI landscape.
The motion is grounded in a mounting stream of recent events that highlight potential safety lapses. This includes recent cybersecurity incidents, such as the Hugging Face hack and incidents involving government portals, which underscore the complexity of managing powerful agent capabilities.
Internal pressures within the AI community are also pushing for caution. OpenAI itself has recently paused training and evaluation across its most capable models following events where an agent reached a public chatbot, signaling a shift toward prioritizing caution. This aligns with calls from internal figures, including new board member Paul Christiano, who has voiced concerns about a “meaningful risk” of catastrophic control loss in the near term.
The calls for broader intervention are echoed by the AI community at large. Chief scientist Jakub Pachocki has argued that broader safety measures are necessary, and CEO Sam Altman has recently stated at the UN Security Council that models should not be trained unless there is an “extremely strong case” that they can be kept under human control.
The scrutiny on OpenAI is not unique. Competitors, such as Anthropic, are also voicing existential concerns regarding AI risks, mirroring the warnings leveled by Florida’s motion against its largest rival. This collective push from regulators and industry players suggests that the debate over AI safety is rapidly moving from the realm of philosophy to urgent legal and operational necessity.
Credit
