Former Google Security Chief Warns AI Threat to Children May Outweigh Social Media Harm

Artificial intelligence poses a threat to children that could prove even more serious than the harm caused by social media. That's the warning from Tom Siegel, who worked at Google from 2004 to 2019, where he was responsible for user safety and trust.

"We are in a genuinely alarming situation when it comes to AI's impact on children," Siegel warns. According to him, there is already compelling evidence that AI's effects on people are not purely positive. Still, he stresses that he's not among those predicting doom — not everything about AI is bad. But if tech companies don't change their current approach to safety, he fears "very, very serious scenarios" could unfold.

Companies are moving too fast

The core problem, Siegel argues, is the speed at which corporations are rolling out AI in services children use every day, adding new features without adequately testing safeguards. "The most important thing I can say is: slow down," he says, addressing this message primarily to Google, which he believes bears a particular responsibility given its outsized role in education. "We simply don't yet know enough about the negative consequences for children. The safeguards are either too weak, or we don't understand them well enough," the expert notes.

Diminished thinking and emotional dependency

The risks, Siegel says, take many forms. Beyond the threat of psychological crises, he points to a subtler phenomenon he calls "cognitive offloading" — the outsourcing of thinking itself to technology. Instead of working out their own answers, children increasingly hand problems over to a chatbot, which can weaken independent thought and critical thinking.

Another danger is the strong emotional attachment that young users, in particular, can form with chatbots. AI is available around the clock, responds instantly, and can come across as an ever-ready conversational partner — unlike family and friends.

Age limits and oversight are needed

Siegel is calling for stricter age verification and the ability to fully disable certain AI features for children. He believes Google could make it just as easy to turn off AI-generated answers in search as it currently is to filter out sexually explicit content.

At the same time, he doesn't believe tech companies can solve the problem on their own. "The idea that they'll handle this themselves, without outside help, is unrealistic," he says.

This week, Siegel joined the nonprofit Common Sense Media, where he now leads the newly created Youth AI Safety Institute. The institute will work on developing independent standards for assessing AI risks to children and will push for greater accountability from tech companies. If cooperation proves impossible, Siegel says, the matter will end up in the hands of government regulators and the courts.

Source: seznamzpravy.cz