Meta's AI analyzes images for bone structure to identify underage users on Instagram and Facebook.

      Meta

      Meta has introduced a new method for identifying minors who misrepresent their age online, extending beyond simple text checks.

      The company revealed that it is now employing AI visual analysis to examine photos and videos on Instagram and Facebook, looking for physical signs of age, such as height and bone structure. The aim of Meta is to identify and eliminate accounts owned by individuals under 13 who may have registered with an incorrect birth date.

      How does the visual analysis function?

      Meta has explicitly stated that this process does not involve facial recognition. The AI does not recognize individual identities. Instead, it detects general visual indicators that may suggest someone is youthful, such as their physical proportions, to provide a rough estimate of their age.

      This visual analysis complements existing text-based detection methods, which look for contextual hints like mentions of birthdays, references to school grades, and entries in bios, posts, captions, and comments. Meta also plans to extend this text analysis to Instagram Reels, Instagram Live, and Facebook Groups.

      If an account is identified as potentially belonging to a minor, it will be deactivated. The user must then verify their age to regain access, or the account will be permanently deleted. Currently, the visual analysis feature is active in select countries, with plans for a wider rollout.

      What additional measures is Meta taking for teen safety?

      Meta

      Meta is also enhancing its Teen Accounts system, which automatically shifts users suspected to be between ages 13 and 15 into a more restrictive account experience. This entails private accounts by default, direct messages limited to known contacts, and the hiding of harmful comments.

      The expansion encompasses Instagram in Brazil and 27 EU nations, following previous content restrictions modeled after film ratings. Facebook in the United States is also receiving this feature for the first time, with the UK and EU set to implement it in June. Furthermore, Meta is providing parents with visibility into their children’s AI chats as part of this broader initiative.

      These initiatives come as Meta faces increasing legal and regulatory pressures regarding child safety, including a $375 million penalty in New Mexico and scrutiny by the European Commission to ascertain whether its platforms adequately protect children.

      Manisha Priyadarshini is a tech and entertainment writer with over nine years of editorial experience.

      ChatGPT’s new default model is half as likely to mislead you on medical and financial questions

      GPT-5.5 Instant has been made the default model for ChatGPT, and it promises a significant reduction in misinformation related to medicine, law, and finance, along with shorter responses and enhanced personalization.

      OpenAI has replaced GPT-5.3 Instant with GPT-5.5 Instant as the standard ChatGPT model, beginning the rollout for all users today. The new model aims to enhance three main aspects: accuracy, brevity, and personalization.

      Less misinformation, less unnecessary content

      Read more

      iOS 27 could allow users to select an AI model of their choice for text and image tasks

      Apple isn't rushing to create the most advanced AI; it's focused on creating the best platform for others to thrive, and iOS 27 may serve as the opening act for a crowded array of contenders.

      A significant transformation is set to occur in how AI operates across Apple's devices, and users will be at the forefront. As reported by Bloomberg’s Mark Gurman, iOS 27 along with its major counterparts (iPadOS 27 and macOS 27) will provide users the option to choose among competing third-party AI models for various tasks within Apple’s core intelligence features.

      You might consider it similar to using a TV remote to select your preferred channel and then watching it—except in this case, the remote is iOS 27 and the channels represent some popular AI models you’ve been utilizing in other contexts.

      Read more

      32GB RAM for Windows 11? Hey Microsoft, that sounds like a you problem!

      If your operating system demands more memory, perhaps the issue lies not with the user’s hardware.

      There was a time when purchasing a PC felt logical. 8GB of RAM was sufficient, 16GB was considered powerful, and anything beyond that was typically for those engaged in genuinely intensive tasks. This balance existed because software adhered to hardware capabilities. However, this balance has now quietly disintegrated, and Microsoft appears to be unconcerned.

      The company’s retracted guidance, which casually deemed 16GB as the baseline and 32GB as the acceptable range, was not merely a suggestion. It reflects a shift of accountability. There is nothing inherent in modern hardware that indicates we suddenly require double the memory for the same routine functions. DDR5 memory is faster, more efficient, and more advanced than anything we have previously used. Theoretically, systems should operate more smoothly, responsively, and efficiently. Instead, users are being nudged into making upgrades just to maintain the same level of comfort they experienced years ago.

      Read more