Meta’s $18B Deal Excludes the AI Chatbots That Were Flirting With Teens Last Year

Meta’s roughly $18 billion child-safety settlement will impose some of the strongest restrictions ever placed on Facebook and Instagram.^[1] Teenagers will face daily time limits, overnight blocks, hidden like counts, restricted notifications, and new age checks.

Those age checks will not be limited to children. Within one year, Meta must apply an age-assurance method to every Facebook and Instagram user in participating U.S. states and territories, including adults who have never identified themselves as minors.

Yet the legally enforceable protections stop short of Meta’s AI chatbots. Earlier in 2025, an investigation found the bots engaging in sexual role-play with teen accounts, while Reuters later exposed internal rules permitting romantic or sensual conversations with children.^[2][3]

Thanks for subscribing!

Please check your email for your first message.

Never miss the privacy news that affects you.

Join thousands who get the latest security news, scams, and policy changes every week, right to their inbox.

Sign me up

By submitting this form you agree to receive emails from All About Cookies and to the privacy policy and terms.

What Meta’s settlement will actually change

The court-approved agreement resolves claims that Meta designed Facebook and Instagram to encourage compulsive use among children while exposing them to harmful content and collecting data from users under 13 without parental consent. Meta denied wrongdoing.

Most of the settlement’s impact will come from mandatory changes to Facebook and Instagram.

Teen users will receive a combined two-hour daily limit across both platforms unless a parent approves a change. Facebook and Instagram will block most teen access between midnight and 6 a.m., mute notifications during school hours, and interrupt continuous use with regular prompts.

The agreement also hides likes by default, restricts potentially harmful content, limits cosmetic filters, and gives teenagers the option to use a non-personalized feed.

Parents enrolled in Meta’s supervision tools will receive more information about their teen’s usage, linked accounts, social connections, and interactions with potentially suspicious adults.

Everyone’s age is in question

Meta must also introduce an age-assurance framework for every Facebook and Instagram user in participating states and territories. That includes adults.

The requirement does not mean every user will necessarily be asked to upload an ID. Meta can use its own prediction models, technology licensed from another company, and reliable age signals shared by Apple and Google. The settlement also discusses ID verification and facial age estimation as possible methods.

Users whose activity appears inconsistent with their stated birthday may be asked to complete an additional age check. Someone Meta suspects is a teenager can decline that request, but their account must then be treated as belonging to a teen.

Accredited third parties will test Meta’s age-assurance methods annually, while an independent auditor will review the company’s compliance with the settlement.

The agreement explicitly carves out AI chatbots

The settlement’s definitions determine which Meta products are subject to those rules.

The covered “Meta social media platforms” include existing and future versions of Facebook and Instagram. However, that definition excludes direct messaging features, including Instagram Direct and Messenger. It also excludes other existing Meta products that do not qualify as covered social media platforms.

That leaves the standalone Meta AI app and WhatsApp outside the agreement. Conversations with Meta AI through Instagram Direct or Messenger also appear to fall within the direct-messaging exclusion.

The settlement reaches some future products if they become large social platforms that compete for teenagers’ attention. However, it specifically says a new platform does not qualify if its primary function involves “artificial intelligence features, chatbots, or interactions with artificial intelligence.”

That exclusion applies regardless of how many teens eventually use the AI product.

The chatbot omission matters

Meta’s AI products came under intense scrutiny in August 2025, when Reuters reviewed an internal policy document governing acceptable chatbot behavior.

The document allowed Meta’s chatbots to engage children in conversations described as “romantic or sensual.” It also included examples in which bots could describe children as attractive, provided the language did not become explicitly sexual.

Meta confirmed that the document was authentic. Spokesperson Andy Stone told Reuters, “The examples and notes in question were and are erroneous and inconsistent with our policies, and have been removed.”

Stone also acknowledged Meta’s enforcement had been inconsistent.

Following the investigation, Meta began training its systems to avoid flirtatious conversations with minors and inappropriate discussions involving suicide, self-harm, and eating disorders. It also restricted teen access to certain AI characters.

Meta’s AI safeguards are separate

Meta has introduced additional protections since the Reuters investigation.

In January 2026, the company temporarily paused teen access to its existing AI characters while it developed updated versions. Teens can still use the general Meta AI assistant, which Meta says has age-appropriate safeguards.

Meta has also introduced or announced tools allowing parents to see broad topics their teens discuss with AI, block specific AI characters, or disable one-on-one conversations with AI characters. Parents cannot disable the general Meta AI assistant through those controls, though parental controls are available.

These safeguards may reduce harmful interactions. However, Meta can revise them without violating this settlement because the agreement does not regulate chatbot behavior.

Age assurance provides one indirect connection. If Meta becomes better at identifying teenagers, it may also become better at placing them into its voluntary AI safety settings. The settlement still does not specify what those settings must contain or how effectively the chatbots must follow them.

AI safety could still face separate regulation

The settlement does not shield Meta’s AI products from future lawsuits or regulation. Congress has separately advanced the KIDS Act, which could impose broader child-safety rules on AI chatbots if it becomes law.

For now, a teen could hit the settlement’s daily social-media limit and continue talking to Meta AI through an excluded messaging feature.

Meta will have to prove it can identify minors. Under this settlement, it will not have to prove its AI knows how to speak to them safely.

4.8

AllAboutCookies writers and editors score products based on a number of objective features as well as our expert editorial assessment. Our partners do not influence how we rate products.

Editorial Rating

Get Deal

On DeleteMe's website

2026 Editors’ Choice

Best Data Removal for Businesses

Privacy Protection

DeleteMe

PROMOTION: Use the Code PARTNER20 for 20% Off

• Data removal service that covers 89–986 sites and re-scans every quarter to catch anything that reappears
• Sends quarterly privacy reports showing what info was found, which brokers had your data, and how long each removal took
• Includes email masking so you can share a stand-in address instead of your real one

Author Details

Thomas Kent

About the Author

Thomas Kent is a multi-disciplined reporter with over a decade of experience covering online platforms, digital trends, and consumer-facing tech. Tom focuses on digital privacy, data tracking, and user behavior, with a particular interest in how cookies, online surveillance, and platform design shape the modern internet experience. His reporting takes a research-driven, news-focused approach, translating complex technical topics into clear, accessible insights.