OpenAI & Anthropic Deploy New AI Tools to Identify Underage Users

For the most part, age checks right now are pretty useless. It just asks if you’re over a certain age, and it’s easy to fake your birthdate. But throughout 2025, we’re seeing companies like Google take active steps to verify a user’s age. Now, both OpenAI and Anthropic are doing their part by announcing new ways of predicting underage users.

OpenAI, Anthropic announce ways of predicting underage users

OpenAI has announced that ChatGPT’s model spec will include four new principles for users under 18. The list includes putting teen safety first, which, according to OpenAI, will be prioritized even if it conflicts with other goals. It will also promote real-world support by encouraging offline relationships.

Teens will also be treated with “warmth and respect,” so instead of condescending answers, teens will be treated like young people, not adults. These are important changes, especially following reports of a couple of incidents where people have died by suicide as a result of interacting with AI models that constantly agreed with them.

On Anthropic’s end, the company already does not allow users under the age of 18 to use its Claude model. It is also rolling out changes that can detect and disable accounts belonging to underage users. This is achieved by detecting certain conversational signals that hint that the user is underage.

What could go wrong?

Yes, there’s no denying that AI is smart and knows many things, but it’s not perfect. We’ve seen instances where AI has been exploited to generate malware. There’s also the common issue of hallucinations, an AI phenomenon where the model generates completely made-up and wrong answers out of thin air.

This means that while we appreciate what OpenAI and Anthropic are doing when it comes to predicting underage users, it’s hard not to be skeptical about its effectiveness. Plus, like we said, Google rolled out similar age-verification tools earlier this year. Unfortunately, it was far from perfect either.

Many users reported that even though they were above a certain age, Google’s systems misidentified them. Users can fix this by submitting documents to prove their age, but it’s an otherwise unnecessary hurdle when all users want to do is search or watch a video online. Will OpenAI and Anthropic have better luck? We’ll have to wait and see.

The post OpenAI & Anthropic Deploy New AI Tools to Identify Underage Users appeared first on Android Headlines.