A teenager lies in a dark bedroom at night looking at a glowing smartphone screen, the face kept in shadow

AI Chatbots and Children: Florida Sues OpenAI, Europe Follows

The first state-led lawsuit against OpenAI turns youth protection for AI chatbots into a liability question. Europe has the rules, enforcement is the open issue.

On 1 June 2026 Florida became the first US state to sue OpenAI and CEO Sam Altman personally, with allegations ranging from missing age verification to behavioural addiction among minors. For you as a decision-maker this is more than a US headline: any company shipping AI chatbots to a broad audience now carries a growing liability and compliance risk. This article explains the lawsuit, the wave of cases behind it, and what the EU AI Act, the Digital Services Act and national youth media protection require from companies.

Summary

Child and youth protection became a concrete risk for AI chatbot providers in 2026. On 1 June 2026 Florida became the first US state to sue OpenAI and CEO Sam Altman personally, alleging the company marketed ChatGPT despite known risks to minors, without effective age verification and parental controls. The lawsuit sits at the peak of a wave: Character.AI banned under-18 chats in late October 2025, Google and Character.AI settled several teen-suicide cases in early January 2026, and the US Federal Trade Commission has been investigating seven companies since September 2025. The numbers explain the pressure: 72 percent of US teens have used AI companions, and about a third prefer them to people for serious conversations. Europe already has the legal framework, from Article 5 of the EU AI Act through the DSA to national youth media protection. For companies this means: youth protection belongs in product design, in age checking, and in a clear escalation path for crises.

Florida sues OpenAI

On 1 June 2026 Florida became the first US state to file suit against OpenAI and its CEO Sam Altman. Attorney General James Uthmeier alleges the company marketed a product whose dangers to minors it knew about. Naming Altman personally sets the case apart from ordinary corporate lawsuits.

1 June 2026
Lawsuit filed
first US state
72 %
teens using AI companions
Common Sense Media, 2025
52 %
regular users
at least a few times a month
7
companies under FTC review
inquiry from Sept. 2025
35M €
maximum fine
or 7 % turnover, AI Act
1 Dec. 2025
KJM enforcement
age verification in Germany

The legal basis is Florida's prohibition on unfair and defective trade practices, and the damages sought could reach into the billions. The trigger was a criminal investigation into chat logs of the perpetrator of an April 2025 shooting at Florida State University. For how legal and ethical duties combine when deploying AI, innobu offers its guide to ethical and legal AI compliance .

The allegations in detail

The lawsuit bundles several allegations that reach beyond the single case. At its core is the claim that OpenAI ignored internal and external safety warnings and put speed before protection. OpenAI counters that minors need significant protection and points to its own safeguards.

What Florida alleges
Missing effective parental controls and age verification
Collecting minors' data without meaningful parental consent
Behavioural addiction and cognitive harm to young users
Dangerous errors downplayed
How OpenAI responds
Points to safeguards and policies for minors
Age prediction to detect under-18 users, stricter mode when unsure
Parents can link accounts, disable features, receive alerts
Self-declared age still remains easy to bypass

Why this goes beyond a single case: The lawsuit ties together product design, data protection and safety. Those same three layers are exactly what European regulators examine. A claim like safe for children is not a check, it is an assertion that has to be proven.

Not an isolated case: the wave of lawsuits

The Florida lawsuit sits at the peak of a longer development. Companion chatbots that simulate emotional closeness have been under pressure since late 2024. Several families sued after teenagers took their lives following intense chat histories.

Late 2024

First suicide lawsuit

A Florida lawsuit names Character.AI in connection with the suicide of a teenage user. More families follow.

September 2025

FTC inquiry

The US Federal Trade Commission asks seven companies, including OpenAI, Meta, Google, Snap and xAI, how they protect children using companion chatbots.

October 2025

Character.AI bans teens

Character.AI bars under-18 chats. In April 2026 it adds mandatory age verification via face scan.

January 2026

Character.AI settlement

Google and Character.AI settle five parallel cases across four US states.

June 2026

Florida sues OpenAI

The first state takes OpenAI and Sam Altman to court. The allegations range from missing age checks to behavioural addiction.

How teenagers use AI companions

The numbers show why this is not a fringe topic. AI companions have become part of many teenagers' daily lives, often as conversation partners for personal matters. That emotional bond is exactly what makes protection so hard.

AI companions are chatbots that take on a social or emotional role in ongoing conversations, such as a friend, confidant or partner. They respond in a personalised way and so encourage a lasting attachment.
A parent and a teenager sit together at a kitchen table in the evening, looking at a smartphone together
Guidance over bans: where parents discuss usage and settings together with teenagers, protection works better than a technical checkbox alone.
72 %
have used AI companions
about 33 %
prefer AI for serious topics
53 %
use AI for homework

Nearly a third of US teens now interact with AI chatbots every day. Usage ranges from homework help and translation to emotional support. That many teenagers would rather share personal worries with an AI than with a person is the real finding behind the lawsuits.

Germany and the EU perspective

Europe does not have to invent these duties, the legal framework is already in place. Unlike in the US, where lawsuits set the standard, several layers interlock in the EU. The open question is enforcement.

Layered diagram of four protection levels for minors using AI chatbots, from the EU level through national rules and provider duties to the parent and user level
Layers of protection for minors using AI chatbots: from EU law through national youth media protection and provider duties to the parent and user level.

Article 5 of the EU AI Act has prohibited manipulative techniques and the exploitation of age-related vulnerabilities since 2 February 2025. The Digital Services Act requires platforms to protect minors, and in July 2025 the European Commission presented a privacy-preserving age verification app using a zero-knowledge process, with integration into the EU wallet planned for 2026. In Germany, the Commission for the Protection of Minors in the Media has been enforcing recognised age verification since 1 December 2025 and calls for AI systems to be written explicitly into the youth media protection treaty.

Key point

Breaches of the Article 5 prohibitions can be fined up to 35 million euros or 7 percent of global annual turnover. For European providers, youth protection is therefore not a voluntary seal of approval but a legal duty with a serious sanction. For how AI security and data protection interlock in a European context, innobu goes deeper in its piece on AI security and data protection .

Challenges and risks

Protection is not a pure win, every measure has side effects. Age verification sits in the tension between youth protection and data protection, and voluntary commitments are often not enough.

Why more protection is needed
Self-declaration and parental checkboxes are easy to bypass
Anthropomorphic bots foster emotional bonds with barely studied effects
Toys and medical devices have clear rules, chatbots so far do not
Where protection hits limits
Age checks by face scan or ID create new data risks
Overly strict filters can block legitimate help in mental distress
Under-18 bans push usage toward unchecked services

The real corporate risk: Running a chatbot without youth protection risks lawsuits, fines and reputational damage, even if minors are not the target group. As soon as children can in fact reach a service, the duties apply. Whoever does not assess the risk still carries it.

What companies should do now

Any company running AI applications that could reach minors should not treat youth protection as optional. The legal and reputational risks rise faster than many roadmaps. Four steps help you stay prepared.

Three colleagues at a glass wall discuss the safety design of a product using a sketch of boxes and arrows
Youth protection belongs in product design: risks to minors are thought through before rollout, not after the first complaint.
  1. Build in safety by design

    Run a risk assessment for minors before a chatbot goes live, not after. Define which topics the bot avoids with teenagers and how it responds to crises.

  2. Age estimation and age-appropriate modes

    Evaluate age estimation methods and offer an age-appropriate mode. When in doubt, the stricter option should be the default rather than relying on self-declaration.

  3. Build crisis routing

    Define clear escalation on signs of self-harm, with referral to support services and a tested, documented response. Test these paths regularly.

  4. Track the legal framework

    Watch the EU AI Act, the DSA and national youth protection law on short review cycles and name owners. That way you spot new duties before they turn into a complaint or a lawsuit.

Key point

Youth protection for AI chatbots is no longer a niche topic for lawyers in 2026, it is part of any serious product development. Building in safety by design, checking age, and escalating crises in a structured way cuts your risk substantially without slowing AI adoption.

Further reading

Frequently asked questions

Why is Florida suing OpenAI? +

On 1 June 2026 Florida became the first US state to sue OpenAI and CEO Sam Altman personally. Attorney General James Uthmeier alleges the company marketed ChatGPT despite known risks to minors, without effective age verification and parental controls. The legal basis is Florida's prohibition on unfair and defective trade practices, and the damages sought could reach into the billions.

How many teenagers use AI companions? +

According to Common Sense Media (July 2025), 72 percent of US teens have used AI companions at least once and 52 percent are regular users. About a third prefer talking to AI over people for serious or personal conversations. A December 2025 study found that nearly a third of US teens interact with AI chatbots daily.

What does the EU AI Act require to protect minors? +

Article 5 of the EU AI Act has prohibited manipulative techniques and the exploitation of age-related vulnerabilities since 2 February 2025. AI systems that deliberately target children or distort their behaviour are banned. Breaches of the Article 5 prohibitions can be fined up to 35 million euros or 7 percent of global annual turnover.

How does Germany regulate AI chatbots for youth protection? +

The Commission for the Protection of Minors in the Media (KJM) has been enforcing recognised age verification systems since 1 December 2025 and calls for AI systems to be written explicitly into the youth media protection treaty and for an independent AI youth protection officer. Generative chatbots that target children or use profiles are treated as high risk.

What should companies running AI chatbots do now? +

Four steps are actionable immediately: first, build in safety by design and run a risk assessment for minors before rollout; second, evaluate age estimation and age-appropriate modes, defaulting to the stricter option when in doubt; third, set up crisis routing with clear escalation on signs of self-harm; fourth, track the EU AI Act, the DSA and national youth protection law on short review cycles and assign clear ownership.