YouTube will begin testing an artificial-intelligence age-verification tool in the United States on Wednesday, expanding technology the company says already works in other regions. The system analyses the types of videos a logged-in viewer watches to estimate whether the user is under 18, overriding the birth date provided at sign-up. If the algorithm identifies a viewer as a minor, YouTube will apply its existing safeguards, including limits on personalised advertising, privacy warnings and prompts to take screen breaks. Accounts incorrectly classified may submit a government ID, credit card or selfie to restore full access, according to a blog post by James Beser, YouTube’s director of product management. Users who are not signed in will continue to face automatic blocks on some material unless they verify their age. The pilot arrives as U.S. lawmakers and courts intensify pressure on online platforms to protect children. In June the Supreme Court upheld a Texas statute requiring pornography websites to verify visitors’ ages, emboldening advocates of stricter oversight. While YouTube and other services are moving ahead with their own checks, civil-liberties groups such as the Electronic Frontier Foundation and the Center for Democracy & Technology warn that broad age-verification mandates could infringe privacy and free-speech rights.