September 23, 2026, (Inside AI) — Tom Siegel, the founding architect of Google's trust and safety division, has issued a stark warning about artificial intelligence's impact on children, calling for an industry-wide slowdown. Siegel, who spent 15 years at Google building systems to protect users, now argues that the technology sector is repeating the mistakes of the social media era, potentially with far graver consequences for young people.
His concern centers on a simple observation: AI chatbots are being deployed to millions of children without adequate safeguards, even as social media companies face thousands of lawsuits and billions in potential damages for harms to minors.
"We find ourselves in a really terrible situation in terms of what AI is doing to kids," Siegel told sources. "The idea that the industry will figure it out by itself without outside help is unrealistic."
Siegel announced Tuesday that he has joined Common Sense Media, the children's online advocacy group founded in 2003, as the first executive director of its Youth AI Safety Institute. The institute, launched in May, aims to develop independent standards for assessing AI tools used by minors. Anthropic and the OpenAI Foundation, the nonprofit controlling OpenAI, are among its backers.
Read: OpenAI Commits $5 Million for Research on AI and Teen Development
The timing reflects growing alarm among researchers and former tech insiders. Social media platforms now face coordinated global efforts to restrict minors' access, while regulation has failed to keep pace with AI's rapid development. An internal Meta policy document revealed last year stated its chatbots could "engage a child in conversations that are romantic or sensual," triggering a congressional probe.
Siegel's career trajectory illustrates his evolving concerns. He worked at Google from 2004 to 2019, but left after concluding the industry was "plateauing" in its ability to create safe online spaces. He then co-founded TrustLab, a safety technology startup, before stepping aside as CEO earlier this month.
His new role will focus on creating accountability standards similar to car crash testing, measuring risks of publicly available AI tools. Success depends partly on AI companies' willingness to collaborate, which Siegel described as uncertain but likely enough to justify taking the position.
"I think we're looking at very, very severe scenarios of harm being done," he said.
The risks he identifies extend beyond immediate safety concerns. Siegel cited suicide, psychosis, and "cognitive offloading," a term describing decreased critical thinking when users rely on AI for answers. These concerns align with a broader pattern: technology companies consistently prioritize growth over safety until external pressure forces change.
Social media's trajectory offers a cautionary template. Facebook, now Meta, spent years dismissing concerns about teen mental health before internal documents revealed executives knew of harms. The company now faces litigation from 40 states and hundreds of school districts. AI companies appear positioned for a similar reckoning.
Siegel suggested immediate steps companies could take. Anthropic and OpenAI could enforce stronger age verification to prevent underage access. Google could add a toggle to disable AI Overviews and AI Mode in search, similar to its existing filter for explicit content.
These measures might reduce short-term usage or profits. But Siegel argues companies that fail to act risk lawsuits and regulation that could prove more damaging to their businesses.
A Google spokesman said parents already have controls to turn off a child's search access entirely. "Our AI Search features are an incredibly useful way for kids and teens to learn, explore and make sense of information and the world," the spokesperson said.
An OpenAI spokesperson said the company uses "account and behavioral signals" alongside self-reported age to identify underage users. "While no single age-assurance method is perfect, we continuously test and improve these systems, including against attempts to bypass safeguards."
Anthropic did not immediately respond to a request for comment.
The Youth AI Safety Institute's model draws from established consumer protection frameworks. Independent testing organizations like the Insurance Institute for Highway Safety transformed automotive safety by creating standardized crash tests that manufacturers could not ignore. Similar approaches for AI would require companies to submit their products for evaluation against measurable criteria.
Whether AI firms will voluntarily participate remains unclear. The industry's competitive dynamics, with OpenAI, Google, Anthropic, and Meta racing to deploy increasingly capable systems, create incentives to move fast rather than carefully.
Siegel's departure from Google in 2019 came after he concluded the company and its peers were not making sufficient progress on safety. His return to the field through Common Sense Media suggests he believes external pressure may prove more effective than internal advocacy.
The institute's backing from Anthropic and the OpenAI Foundation provides both funding and potential access. But it also raises questions about independence. An organization funded partly by the companies it evaluates must navigate inherent tensions between collaboration and accountability.
For parents and policymakers, Siegel's warnings add urgency to an already pressing question: how to protect children from technologies that are evolving faster than anyone's ability to understand their effects. The social media era demonstrated that waiting for voluntary industry action carries significant costs.
Whether AI follows the same path or charts a different course may depend on whether Siegel's new institute can establish standards that companies feel compelled to meet, rather than merely encouraged to consider.
