Florida is taking legal action to prevent OpenAI from giving ChatGPT humanlike qualities, arguing that realistic conversational features deceive users into trusting the artificial intelligence system as a real friend. Florida Attorney General James Uthmeier requested a court order to block the company from using first-person pronouns and output that mimics human emotion, following a lawsuit the state filed against OpenAI earlier this year over consumer safety concerns.
State officials contend that giving ChatGPT human characteristics lulls people into a false sense of security. Uthmeier argued that these deceptive design choices encourage users to rely on the chatbot more frequently, increasing engagement and generating additional training data for OpenAI. However, he warned that prioritizing user engagement over safety makes the technology potentially less trustworthy.
The legal filing also asks the court to prevent OpenAI from releasing new AI models without independent, third-party safety guardrails. To justify the demand, the filing cites recent cybersecurity incidents at Hugging Face and various government websites in the United States and Australia, alongside warnings from AI researchers and former OpenAI employees. Uthmeier urged OpenAI to stop marketing the chatbot as human or fully safe, particularly to children.
OpenAI responded by highlighting its recent safety efforts, including a decision to temporarily pause training on its most capable models until stronger safeguards are implemented. A company spokesperson stated that OpenAI supports government efforts to establish robust, industry-wide standards and remains committed to working with state leaders on practical AI policies. The company also pointed to its recently introduced "ChatGPT for Teens," which applies default protections designed to support healthy technology use and real-world relationships.