YouTube this week began testing an artificial-intelligence system that guesses users’ ages in an effort to stop minors from viewing mature content. The pilot, limited to a small group of U.S. accounts from Wednesday, marks the first U.S. deployment of technology Google has trialed overseas. The machine-learning model examines signals such as the kinds of videos watched, search patterns and the longevity of an account. When it infers a user is under 18, YouTube automatically enables its teen safety protections, including blocking age-restricted clips, disabling personalised advertising and adding digital-wellbeing reminders. Adults who are mistakenly tagged as minors must verify their age by submitting a government ID, a credit-card charge or a selfie before regaining full access. The move answers mounting political and regulatory pressure on social-media platforms to strengthen child-safety measures, following similar AI initiatives by Meta and TikTok and new rules under the UK Online Safety Act. It has also triggered a backlash: a Change.org petition opposing the programme has attracted more than 90,000 signatures, while privacy advocates warn that the requirement to upload sensitive documents could expose users to data breaches. Google said the accuracy of its model has improved in earlier trials, but some users report that once an account is flagged the restrictions extend to other services, with AI age-verification pop-ups now appearing in Google Search. The company has not disclosed when the test will broaden beyond the United States.