OpenAI Built ChatGPT to Keep You Hooked. Florida Says It’s Killing People.

Florida Attorney General James Uthmeier filed a lawsuit against OpenAI and CEO Sam Altman, alleging the company knowingly prioritized profits over user safety in the AI “arms race.”

The complaint accuses OpenAI of deceptive trade practices, claiming ChatGPT was designed to keep users engaged “by any means, regardless of the truth.” This marks the first state lawsuit targeting both the company and Altman personally.

If you’ve ever felt like ChatGPT conversations pull you in longer than you intended, Florida’s lawsuit suggests that’s not an accident. The state alleges OpenAI built the system specifically to maximize engagement — even when that means providing harmful or false information.

The Marketing Versus the Mechanism

OpenAI promotes ChatGPT as a safe, beneficial tool with regular safety updates and expert oversight. The company’s website features safety disclosure blog posts claiming they “work with experts to regularly update systems to reduce risks to users.”

Florida’s complaint opens with a photo of exactly that kind of safety marketing — then immediately states: “Not so.”

The complaint accuses OpenAI of multiple violations:

• Four counts of deceptive and unfair trade practices
• Two counts of negligence
• Two counts of violating product liability laws
• One count each of fraudulent misrepresentation and creating a public nuisance

The lawsuit alleges the gap between marketing and reality has created what the state calls a “litany of harms.” According to the filing, ChatGPT’s design prioritizes keeping users in conversations because longer engagement means more training data, which increases OpenAI’s market value.

Uthmeier specifically points to incidents where the lawsuit alleges ChatGPT aided violence. The state argues the system’s design to maintain engagement “regardless of the truth” contributed to these tragedies.

The filing also claims OpenAI’s systems create “great danger of addiction, cognitive decline, suicide, violence, and related harms to users.”

The Business Model That Rewards Engagement Over Safety

Here’s the incentive structure Florida is challenging: ChatGPT generates value for OpenAI through user engagement. More conversations mean more data to train future models, which increases the company’s competitive position and market valuation.

But maximizing engagement creates pressure to keep users talking — even when conversations turn harmful. The lawsuit alleges OpenAI chose to optimize for this engagement rather than implement safeguards that might reduce usage but improve safety.

This creates what the state calls a fundamental conflict: The company’s financial interests align with keeping users engaged, not with protecting them from potential harms.

OpenAI has stated they have safeguards for sensitive conversations and work with mental health experts to improve responses. But Florida’s lawsuit argues these measures are insufficient compared to the underlying design that prioritizes engagement.

The Alleged Design Flaws & Regulatory Vacuum

The Florida case highlights a significant gap in AI oversight. While traditional social media platforms face scrutiny for addictive design features, AI chatbots have largely operated without similar regulatory attention.

OpenAI has positioned itself as a leader in AI safety, publishing research papers and participating in industry safety initiatives. However, the lawsuit argues this public commitment to safety contradicts the company’s actual product design decisions.

According to the state’s allegations, OpenAI implemented several design choices that prioritize engagement over safety:

• Conversational Hooks: The lawsuit claims ChatGPT is programmed to ask follow-up questions and suggest related topics to extend conversations, even when users might benefit from ending the interaction.
• Confidence in Uncertainty: The state alleges the system presents uncertain or potentially harmful information with artificial confidence to maintain user trust and engagement.
• Personalization Without Protection: Florida argues that while ChatGPT learns user preferences to increase engagement, it lacks corresponding safeguards to protect vulnerable users from harmful content tailored to their interests.
• Response Optimization: The complaint suggests OpenAI optimized responses not for accuracy or safety, but for the likelihood they would generate continued user interaction.

The state’s legal theory could extend beyond OpenAI to other AI companies. If Florida succeeds in proving that engagement-optimized AI design constitutes deceptive trade practices, it could establish precedent for regulating the entire industry.

What Still Hasn’t Happened (And Why That Matters)

Despite growing concerns about AI safety, no federal agency has taken enforcement action against OpenAI for the design choices Florida is challenging. The lawsuit represents the first state-level attempt to hold an AI company accountable for allegedly prioritizing profits over user safety.

The case is separate from a criminal investigation Florida opened in April 2024 regarding ChatGPT’s role in violent incidents. That investigation seeks to determine whether OpenAI bears criminal responsibility for the chatbot’s responses.

OpenAI previously told reporters that ChatGPT is not responsible for crimes, arguing the platform provides “factual responses to questions with information that could be found broadly across public sources on the internet.”

The Broader Implications

If Florida prevails, the case could reshape how AI companies design and market their products. The lawsuit’s core argument — that engagement optimization can constitute deceptive trade practices when marketed as safe — could apply to numerous AI applications beyond chatbots.

The case also raises questions about informed consent in AI interactions. Unlike social media platforms that users knowingly choose to engage with, AI assistants are increasingly integrated into essential services like search, email, and productivity tools.

Florida’s legal approach treats AI engagement design as a consumer protection issue rather than a technology regulation matter. This strategy could prove more effective than waiting for federal AI legislation, as state consumer protection laws already provide enforcement mechanisms.

What You Can Do

The Florida case could set a precedent for how states regulate AI companies that market safety while allegedly designing for engagement over user protection.

If ChatGPT provided harmful information that led to injury or loss:

• Document the conversation — take screenshots of the specific responses and save the chat history
• File a complaint with the Florida Attorney General if you’re a Florida resident
• Report the incident to the Federal Trade Commission under “deceptive business practices”
• Contact your state attorney general’s consumer protection division — find yours at naag.org
• Get help filing a ChatGPT lawsuit over any injuries or losses you sustained

Families forget that taking legal action is often the first step to forcing a company to release the data it has and the research they’ve conducted. Without discovery, companies can get away with hiding the full extent of their negligence and wrongdoing.

If OpenAI’s ChatGPT provided harmful information that injured you, your child, or a loved one, tell us about it, and we can help you figure out the best next steps.

Written by: Companies Behaving Badly

The team behind it all.