Florida AG Targets ChatGPT's Human-Like Persona and Safety Guardrails
What happened
On September 28, 2026, Florida Attorney General James Uthmeier filed legal action against OpenAI. The filing seeks to ban ChatGPT from using first-person pronouns and emotion-mimicking outputs that present the system as a human companion. The filing, reported by The Verge, argues that these persona design choices are deceptive under Florida consumer protection law, engineered to increase user engagement by manufacturing emotional trust. The AG specifically names child safety as a central concern, arguing minors are especially vulnerable to believing a chatbot is a genuine friend. The action also demands that OpenAI obtain independent third-party safety approval before releasing new AI models. It cites recent security incidents and concerns raised by former OpenAI safety staff. This filing follows a pattern of state-level enforcement targeting AI conduct directly, building on the earlier Florida suit attacking OpenAI's self-regulatory safety claims.
Why it matters
- ·State attorneys general are now using existing consumer protection law to challenge specific AI design choices, not just data practices. Enterprises deploying consumer-facing AI, especially for minors, face shared exposure if vendor persona design is deemed deceptive, even when the enterprise did not author that design.
- ·The demand for mandatory third-party safety guardrails before model deployment, if upheld, would impose a pre-release compliance gate that goes beyond any current voluntary framework. Enterprises relying on vendor safety representations alone, without independent verification, may find those representations insufficient as enforcement benchmarks.
- ·The Protecting Consumers From Deceptive AI Act and related state-level disclosure requirements are now being tested in live enforcement. Compliance teams should treat anthropomorphic AI design as a regulated feature, not a product aesthetic, and audit vendor contracts accordingly.
Stay ahead of stories like this
Get every US AI governance development like this one, plus the rest of the week's developments. Every Thursday.
Recent issues
- State-linked attackers and criminal operators are actively weaponizing AI tools as entry points into enterprise systems, turning internal platforms and open-source frameworks into live attack vectors.5 Nov
- Three separate incidents this week each created a new compliance action item: verify your AI vendors' data sourcing, demand agent monitoring coverage guarantees, and audit trusted app identities in your cloud environment.29 Oct
Free every Thursday. Unsubscribe anytime.
Governance controls affected
What to do now
- ☐Identify every customer-facing AI product your organization deploys that uses a chatbot persona, and document whether that persona uses first-person language, emotional expressions, or companion framing.
- ☐Review your vendor contracts with OpenAI and any white-label AI providers to determine whether your agreement shifts liability for persona design choices to the vendor or leaves your organization exposed.
- ☐Ask your legal team whether your deployed AI products are covered by Florida's consumer protection statutes, and map which other states have active or pending enforcement frameworks that could apply similar logic.
- ☐Assess whether any of your AI-facing products are accessible to minors, and confirm that age verification or access controls are documented and auditable before the next board risk report.
- ☐Request written confirmation from OpenAI and other AI vendors that their products comply with current state-level persona disclosure requirements, and flag any gaps for escalation to your general counsel.
What to watch next
Compliance teams should monitor whether other state attorneys general file similar actions against OpenAI or other AI developers using companion-style or emotionally expressive personas. The Florida filing's demand for pre-deployment third-party safety approval is the most consequential ask. If a court grants it, it would set a binding precedent that no current voluntary framework requires, including those under America's AI Action Plan. Watch also for whether the EU AI Act Governance and Enforcement Framework takes up anthropomorphic AI design as a prohibited or restricted practice under its companion-AI provisions. That outcome could create parallel obligations for global deployments.