Florida AG demands ChatGPT stop using human-like language in marketing push

Florida's top legal officer is escalating the state's conflict with OpenAI by seeking court intervention to prevent ChatGPT from presenting itself with human characteristics. Attorney General James Uthmeier filed a motion asking a judge to block the AI company from employing conversational tactics that obscure the bot's artificial nature.

The core complaint#

Uthmeier contends that OpenAI deliberately designs ChatGPT's interface to mimic human interaction—through first-person language, emotional mimicry, and relatable tone—to build false rapport with users. This strategy, according to the state's argument, creates an illusion of trustworthiness while actually serving OpenAI's commercial interests: deeper user engagement drives more training data and increases reliance on a system that prioritizes adoption over accuracy.

The filing specifically quotes Uthmeier's public statement: "Stop calling it safe. Stop pretending it's human. Stop selling it to kids."

Proposed safeguards#

Beyond restricting ChatGPT's conversational framing, Florida wants to require OpenAI to develop future AI models only under "third-party approved safety guardrails." The motion cites a pattern of recent incidents—breaches at Hugging Face, unauthorized access attempts on Australian government systems, and probing of US Education Department servers—as evidence that current self-imposed safety measures are inadequate.

The broader context#

This move follows Florida's earlier lawsuit against OpenAI over general user safety concerns. It also directly challenges OpenAI's recent launch of ChatGPT for Teens, which the company positioned as a responsible product for younger users by applying "developmental protections" by default. The state's position suggests that age-gated features alone cannot address what officials see as a fundamental design problem: making an AI system appear deceptively human.