Teen chatbot use exposes gap in social media age bans

Teenage use of AI chatbots is expanding while governments move to curb underage access to social media, creating a new regulatory gap for technology platforms and policymakers. Roughly half of U.S. teenagers use tools such as ChatGPT, Microsoft Copilot and Character.AI for schoolwork, information or entertainment, according to Pew Research Center.

Digital safety specialists told CNBC that new child protection measures aimed at social media often do little to address AI systems that can hold sustained, personalised conversations with minors. The concern is that chatbots may reproduce some of the dependency patterns associated with social media while operating under less developed rules.

Regulators target social media first

Australia became the first country to legally enforce a teen social media ban in December, CNBC reported, turning the policy into a test case for other governments. The U.K., Spain, France, Greece and Canada have since pursued similar measures, while state-level restrictions have gained attention in the U.S.

Those policies reflect years of concern over social media design features, including infinite scrolling and recommendation systems that can keep users engaged for long periods. Earlier this year, companies including Meta, which owns Facebook, Instagram and Threads, and Google’s YouTube were found negligent over failures to warn users adequately about risks on their platforms, CNBC reported. The alleged harms included addictive product features and body dysmorphia.

AI chatbots work differently from social feeds. Rather than presenting a stream of posts, they respond directly to user prompts and can simulate supportive, knowledgeable or emotionally affirming exchanges. CNBC cited evidence that some teenagers use chatbots as substitutes for real-life friendships and relationships, alongside research describing addiction-related patterns.

AI companions receive narrower treatment

Kaitlyn Regehr, associate professor of digital humanities at University College London, told CNBC that governments spent years reacting to social media harms and risk repeating that delay with AI products made available to children. She said existing policy discussions often concentrate on the most extreme chatbot risks while paying less attention to broader dependency and cognitive effects.

The U.K.’s approach has included a limited reference to restricting under-18s from AI romantic companion chatbots designed to support sexual relationships or roleplay, according to CNBC. In the U.S., the House has passed the KIDS Act, which would restrict AI chatbot interactions with children, though the measure still requires Senate approval.

Sonia Livingstone, a London School of Economics professor focused on children’s digital rights and online safety, told CNBC that AI safety is not necessarily absent from policy discussions, but investment in AI appears to be taking priority. She said regulation is still often treated as a constraint on innovation rather than a route to commercially viable and trusted products.

The tension is visible in the U.K., where the government promoted billions of pounds of AI investment and its ambition to make Britain an AI superpower shortly before presenting a major social media ban for under-16s, CNBC reported.

For technology companies, the debate adds a child-safety dimension to the wider AI buildout. For governments, it raises a policy question already familiar from the social media era: whether regulation can keep pace with products that reach children before their risks are fully understood.

This story draws on original reporting from CNBC.