Google Tests AI-Powered Age Estimation for Enhanced User Protection

Google Tests ML-Based Age Estimation for Minors in the U.S.

Disclaimer: This content generated by AI & may have errors or hallucinations. Edit before use. Read our Terms of use Google is set to test a Machine Learning (ML) system for age estimation in the United States. According to the company's blog post, "New Digital Protections for Kids, Teens, and Parents," the software aims to predict whether a user is a minor to provide them with more "age-appropriate experiences." This development was first announced in YouTube CEO Neal Mohan's annual letter, where he stated that the platform would leverage ML to "distinguish between younger viewers and adults." Governments worldwide are pushing for stricter age verification rules, as exemplified by Australia launching an age verification trial to evaluate the effectiveness of current technologies in determining a user's age and regulating access to age-restricted online content. Later, the country banned social media for children under 16, prompting similar deliberations in the U.K. Beyond Australia, India's Digital Personal Data Protection (DPDP) Act 2023, the US' Children's Online Privacy Protection Act (COPPA) 1998, and the UK's Age-Appropriate Design Code regulate data collection for minors. To comply with these legislations, social media platforms like TikTok, Meta, Google, and Snap have undergone several regulatory changes. Notably, in 2020, YouTube announced that if its systems fail to verify a user's age, they will request a valid ID or credit card for age confirmation. Despite various online platforms like YouTube, Instagram, TikTok, and Facebook implementing age verification technologies, their effectiveness remains questionable. According to a 2023 report by the French Data Protection Commission, CNIL, no age verification solution fully meets privacy-friendly requirements. For example, YouTube's use of payment card validation for age verification has proven ineffective, as many minors today have access to payment cards, enabling them to bypass restrictions. Additionally, mechanisms such as identity document verification, facial analysis, and government-provided tools have failed to meet universal age verification standards. Experts at MediaNama's event on the draft DPDP Rules 2025 pointed out that age verification is not a "magic solution" to children's online safety concerns. Many speakers recommended separate child safety legislation, such as an age-appropriate design code, as a more comprehensive solution. YouTube's age estimation model relies on analyzing existing user data, including visited sites and consumed content, as reported by The Verge. In the US, the platform can legally conduct behavioral monitoring if it obtains verifiable parental consent under COPPA 1998. However, extending this feature to India presents challenges. Under India's DPDP Act 2023, platforms cannot track or monitor children's behavior unless the Central Government determines that they process data in a "verifiably safe manner." YouTube's Government Affairs and Public Policy Manager, Hui Lin Tan, contested this restriction at a 2023 discussion by The Quantum Hub. She argued that behavioral monitoring is integral to YouTube's age-appropriate recommendation system and that limiting it could lead to unintended consequences, such as exposing minors to harmful content or making appropriate content less accessible to them. In October 2024, as US lawmakers proposed legislation to verify the age of internet users, Google released its "Legislative Framework to Protect Children and Teens Online" report. While acknowledging the need to determine user age for age-appropriate experiences, Google proposed alternative verification approaches that preserve user anonymity. The company recommended parental control tools, risk-based impact assessments, and bans on personalised ads, among other measures. Additionally, Google stated that "data-intrusive" methods, such as government ID verification, should be restricted to high-risk services like alcohol or gambling platforms. While Google and YouTube spoke briefly about the initiative, there was a lack of clarity on how the model would work. MediaNama reached out to Google with a few questions and will update the story as soon as we hear back from them. Alongside the machine-learning-based age estimation, Google announced several updates to its parental control mechanism, 'Family Link.' The new features include "School Time," which allows parents to limit a child's device access during school hours, and 'tap to pay' via Google Wallet for Android devices. Additionally, the company is expanding access to educational features like Learn About and NotebookLM for teens.

Google Will Use AI to Determine if You're Lying About Your Age

Samantha Kelly is a freelance writer with a focus on consumer technology, AI, social media, Big Tech, emerging trends and how they impact our everyday lives. Her work has been featured on CNN, NBC, NPR, the BBC, Mashable and more. Google said it will begin testing a machine learning-powered age estimation model to help determine if a user is under the age of 18. The company said in a blog post that the model, expected to launch later this year, intends to provide "more age-appropriate experiences" for young users. It will roll out first in the US and expand to more countries over time. The announcement comes as tech companies face mounting pressure from lawmakers to make platforms safer for young users. Last year, Meta introduced a similar AI model that helps determine if a user is under the age of 18. Google told CNET the estimation model will use machine learning to make sense of various data points already associated with a user's account, such as what they're searching for, the categories of videos they've watched and how long they've had an account. These signals can help the model better determine if someone may be lying about their age. If the model incorrectly flags an account for being under the age of 18, the user can provide age verification in the form of a government-issued ID, credit card or selfie. Google said it is also looking into other ways people will be able to verify their age. This method has already been used by other companies, such as Meta, and in a variety of markets, including Europe. YouTube CEO Neal Mohan mentioned the machine learning tool in the company's published priorities for 2025, noting the company plans to continue to "responsibly harness the power of AI to enhance the YouTube experience for everyone." The company is also adding a handful of safeguards to underage accounts, such as filtering or blurring explicit results in search and restricting sensitive ad and other content on YouTube that may not be appropriate. Meanwhile, a new feature called School Time to Android phones and tablets will allow parents to limit or restrict app access during school hours. Other upcoming changes include making it easier for parents to manage their kids' online experiences across Android and Chrome devices via Google Family Link, and the ability to "tap to pay" for Android phones with parental consent.

Google to test using AI to determine users' ages

Google chief executive Sundar Pichai speaks during the tech titan's annual I/O developers conference on May 14, 2024, in Mountain View, California. Google will start using artificial intelligence to determine whether users are age appropriate for its products, the company said Wednesday. Google announced the new technique for determining users' ages as part of a blog focused on "New digital protections for kids, teens and parents." The automation will be used across Google products, including YouTube, a spokesperson confirmed. Google has billions of users across its properties and users designated as under the age of 18 have restrictions to some Google services. "This year we'll begin testing a machine learning-based age estimation model in the U.S.," wrote Jenn Fitzpatrick, SVP of Google's "Core" Technology team, in the blog post. The Core unit is responsible for building the technical foundation behind the company's flagship products and for protecting users' online safety. "This model helps us estimate whether a user is over or under 18 so that we can apply protections to help provide more age-appropriate experiences," Fitzpatrick wrote. The latest AI move also comes as lawmakers pressure online platforms to create more provisions around child safety. The company said it will bring its AI-based age estimations to more countries over time. Meta rolled out similar features that uses AI to determine that someone may be lying about their age in September. Google, and others within the tech industry, have been ramping their reliance on AI for various tasks and products. Using AI for age-related content represents the latest AI front for Google. The new initiative by Google's "Core" team comes despite the company reorganization that unit last year, laying off hundreds of employees and moving some roles to India and Mexico, CNBC reported at the time.