Meta’s AI age check: How it detects kids without facial recognition

Meta has introduced a new AI-driven system to identify underage users across its platforms, marking a significant shift in how social media companies enforce age restrictions. The move comes amid growing global scrutiny over child safety online, with the company aiming to proactively detect and act against accounts that may belong to users below the permitted age.

Meta Platforms has introduced an artificial intelligence-based system to identify users under the age of 13 on Instagram and Facebook. The move is part of a broader effort to strengthen age verification and improve child safety across its platforms.

Unlike earlier systems that relied mainly on self-declared age, this tool actively analyses user activity and content to detect whether an account may belong to a minor.

How does the AI detect underage users?

The system uses a mix of visual and contextual analysis. It scans photos and videos for general physical cues such as height and bone structure, while also examining captions, bios, comments, and interaction patterns for signals that may indicate a user is underage.

Meta has stated that the tool does not use facial recognition and does not identify specific individuals. Instead, it detects patterns and assigns a probability that an account belongs to someone below the minimum age limit. This approach is similar to technologies used by Yoti and k-ID.

What happens if an account is flagged?

If the AI system determines that an account may belong to a user under 13, the account can be deactivated. To regain access, the user must complete an age verification process.

This represents a stricter enforcement approach compared to previous measures, signalling Meta’s shift towards proactive detection rather than relying solely on user honesty.

Meta is also expanding its “Teen Accounts” feature for users aged 13 to 17. These accounts come with built-in safety restrictions, including:

• Limited messaging from unknown users
• Stronger content filtering
• Restrictions on features such as livestreaming for younger teens

While Instagram introduced these safeguards earlier, Facebook is now implementing them in the United States, with plans to expand to other regions such as the United Kingdom and the European Union.

Why is Meta introducing this now?

The rollout comes amid increasing legal scrutiny over child safety on social media platforms. A recent jury ruling in New Mexico found Meta liable for misleading users about safety and failing to adequately protect children, resulting in significant financial penalties.

In this context, the introduction of AI-driven age detection appears to be part of a broader response to regulatory pressure and growing concerns about online harm.

While the system aims to enhance child safety, it also raises important questions. These include the accuracy of AI-based age estimation, the possibility of legitimate users being incorrectly flagged, and concerns around privacy and data usage.

As the feature is currently being tested in select regions, its effectiveness and wider impact will be closely monitored as Meta expands the rollout globally.