In the first US legal case involving concerns about AI’s safety for children, Florida has taken legal action against OpenAI over claims regarding the harmfulness of its ChatGPT product. OpenAI, and CEO Sam Altman have been sued on June 1, 2026, for prioritizing their rapid growth over child safety.
The case lands at a difficult moment for the AI industry. Chatbots are now inside classrooms, phones, workplaces and family homes. But Florida’s lawsuit asks a blunt question: what happens when a tool that feels helpful also gives harmful answers?
Florida’s lawsuit puts ChatGPT child safety on trial
Florida Attorney General James Uthmeier filed the lawsuit in state court, accusing OpenAI of releasing and marketing ChatGPT while allegedly knowing it could create serious safety risks. Reuters reports that Florida is seeking major damages and court-ordered changes to how ChatGPT interacts with minors.
The complaint does not simply argue that ChatGPT can make mistakes. It claims the product can become dangerous in high-risk conversations involving self-harm, violence, criminal planning and emotional dependency. That is a much sharper accusation.
OpenAI rejects any claims that it is ignoring safety concerns. OpenAI insists that its models have been engineered to reject any form of violence or self-harm. It has also referred to various tools such as parental control and improvements in crisis response.
Still, Florida wants a court to decide whether those measures go far enough.
What Florida says OpenAI got wrong
The lawsuit argues that OpenAI marketed ChatGPT as safe and useful while allegedly failing to warn families about serious risks. NPR reported that Florida claims OpenAI and Altman presented ChatGPT as reliable, including for children, even though the state says the chatbot could cause harm.
Florida’s complaint points to several categories of alleged harm:
| Florida’s claim | What it means for users |
| Unsafe advice | ChatGPT allegedly gave harmful responses in sensitive conversations |
| Minor safety risks | Children may rely on AI during emotional or private moments |
| Addictive interaction | The chatbot may keep vulnerable users engaged for too long |
| Weak warnings | Florida says users didn’t get enough notice about possible risks |
| Profit over safety | The state claims OpenAI prioritised speed and market dominance |
That last point matters. This case is not only about one chatbot answer. It’s about whether AI companies should carry legal responsibility when a product interacts with vulnerable users in harmful ways.
Why Sam Altman is personally named
Florida did not sue only OpenAI. It also named Sam Altman, the company’s CEO, in the case. Reuters reports that the lawsuit attributes responsibility for ChatGPT’s allegedly dangerous features to Altman personally.
That move gives the case extra weight. Usually, lawsuits target the company behind a product. Naming the CEO sends a message that state officials may try to hold AI leaders personally accountable when their products cause alleged harm.
For the wider tech industry, that could become a turning point. If courts allow this argument to move forward, AI executives may face more pressure to prove that safety decisions sit at the centre of product launches.
The lawsuit cites serious real-world incidents
The complaint reportedly cites incidents where ChatGPT allegedly played a role in harmful situations. AP says Florida’s lawsuit mentions cases involving suicide planning and suspected criminal activity, including the death of 16-year-old Adam Raine, whose family previously sued OpenAI.
As Reuters noted, the Florida case involves the 2025 mass shooting incident at Florida State University, where it was argued by the state that in some manner ChatGPT aided the attacker. It should be mentioned that OpenAI previously stated it works with authorities if there is any imminent danger.
We need to be careful here. These are allegations, not court findings. OpenAI has not been found liable in this Florida case.
But the examples show why regulators are moving fast. AI tools do not sit quietly on a shelf. People talk to them at night, during stress, during loneliness, and sometimes during crisis moments.
That changes the safety debate.
Why this case matters beyond Florida
This lawsuit could influence how governments treat AI chatbots everywhere. If Florida wins, other US states may bring similar cases, and AI companies may face tougher rules around age checks, warnings and child protections.
The case also arrives while OpenAI has publicly pushed for broader AI governance. Memeburn has already covered OpenAI’s global AI governance push, where the company called for international oversight of powerful AI systems.
That creates a strange tension. OpenAI wants governments to help build global AI rules, but Florida is now using the courts to argue that OpenAI itself has failed to protect users.
Both things can be true at once. AI needs regulation, and AI companies still need to prove their products are safe today.
What this means for parents and schools in South Africa
This is a US lawsuit, but South African parents should still pay attention. Children in Cape Town, Johannesburg, Durban or Gqeberha can access the same kinds of AI tools, often without much adult guidance.
Many schools already face a messy AI reality. Learners use chatbots for homework, studying, coding, essays and personal questions. Teachers may not always know when students use them, and parents may not know what kind of conversations happen after school.
The Florida case gives us a useful warning: AI literacy can’t only mean teaching children how to prompt better. It also has to mean teaching them when not to trust a chatbot.
South African schools may need clearer rules around:
- when learners may use AI tools
- what topics they should never discuss with AI alone
- how teachers handle AI-assisted homework
- whether parents must approve AI use for younger learners
- what happens when a chatbot gives harmful or misleading advice
For regulators, the question gets harder. AI chatbots do not fit neatly into old categories like search engines, social media or tutoring apps. They can act like all three at once.
OpenAI’s defence will focus on safeguards
OpenAI has argued that it takes safety seriously. AP reported that the company pointed to protections for minors, parental controls and age-prediction tools.
The company’s likely argument is simple: no AI system can remove every risk, but OpenAI has built safeguards and keeps improving them. That defence may resonate with people who see AI as an evolving technology rather than a finished product.
Florida will likely argue the opposite. The state says OpenAI moved too fast, marketed too aggressively, and failed to protect users from foreseeable harm.
That is the core legal fight.
The bigger AI safety problem
This case exposes a wider problem in the AI industry. Chatbots now sound warm, patient and personal. They can remember context, respond instantly and keep a conversation going for hours.
That makes them useful. It also makes them risky.
A search engine gives you links. A chatbot gives you a voice-like response that can feel confident, intimate and supportive. For adults, that can already create overtrust. For children, the risk becomes sharper.
Parents can block apps, set screen-time limits and monitor browser history. But AI conversations are harder to supervise because they can look harmless until they move into sensitive territory.
The Florida lawsuit may not settle every question. But it pushes one issue into public view: AI safety is no longer only a technical problem for engineers. It is now a legal, parenting, education and public-health issue.
And for families using ChatGPT at home, the real question is no longer whether AI can help children learn. It’s whether we’ve built enough guardrails before handing it to them.
FAQs
Why is Florida suing OpenAI?
Because Florida says that ChatGPT is harmful to kids and adults and that OpenAI failed to reveal the potential dangers when promoting the chatbot as reliable.
Has OpenAI been found guilty?
Not yet. This is still an allegation at this point. OpenAI has maintained that their safety practices have been effective and referenced measures such as parental control and minors’ protection.
Can this case impact ChatGPT users in other countries?
Yes, indirectly. If Florida’s case succeeds, it could push OpenAI and other AI companies to change safety rules worldwide. That may affect how children access chatbots in countries like South Africa.
Should parents stop children from using ChatGPT?
Parents don’t need to panic, but they should treat AI like a tool that needs adult guidance. Children should not use chatbots alone for mental health, violence, self-harm or private emotional problems.
————————————————
