CVCOVANewsComputing, Opportunities, Value & Analysis
Cybersecurity & Privacy4 min read

Instagram will use AI worldwide to detect teens who lie about their age: what we still don't know

Meta has expanded worldwide the system Instagram uses to estimate its users' ages and move teenagers into more restricted accounts. The company has explained little about what data it analyzes, how accurate the tool is, or how to correct a mistake.

By COVA News newsroom
Instagram will use AI worldwide to detect teens who lie about their age: what we still don't know

A 15-year-old girl opens Instagram with a made-up date of birth that lets her pass as an adult. Until recently, in much of the world, that small deception was enough. That is starting to change: on November 9, 2026, Meta announced that it is rolling out worldwide the artificial intelligence system Instagram uses to try to infer the real age of the people using it. If the tool suspects that a teenager is behind an “adult” account, it moves the account into Teen Accounts, which limit who users can talk to, what content they see and how much time they spend in the app.

For mothers, fathers and teachers in Latin America, the news cuts both ways. It promises more protection for minors, but it relies on a system that observes the behavior of millions of people and about which Meta has said very little.

According to reports from several outlets, these special accounts cover young people ages 13 to 17. Users under 16 need permission from a parent or guardian to change their settings, and if the system believes the account holder is under 13, the account is deactivated until they prove their age. Al Jazeera also mentions a limit of two hours of use per day, a detail that does not appear confirmed in the Meta information that could be reviewed.

The tool is not new. Meta has been using artificial intelligence since 2025 to detect people who lie about their age, and the technology was already operating in the United States, the European Union and Brazil. One outlet places its arrival in the 27 European countries and Brazil in May 2026. The global expansion comes three months after the company signed an agreement in August with 52 U.S. attorneys general that includes strengthening the detection of minors. It also coincides with a global debate over who should verify users' ages, in which Meta argues that the task should fall to app stores.

An AI that estimates age, but by looking at what?

The big question is what data the machine examines, and here the accounts do not match. On November 9, Engadget corrected its initial story, which had referred to the analysis of bone structure in photos and to mentions of school grades or birthdays. Other sites, such as Android Ayuda, analyst Pasquale Pillitteri's site and the privacy consultancy Captain Compliance, still describe visual or text signals and attribute them to a Meta post from May 2026. On the other hand, according to Infobae, Meta stated in May that this is not facial recognition but an approximate estimate, and Captain Compliance cites a July 31 correction stating that no biometric data — that is, physical traits unique to each person — is used. The November announcement does not detail any of these signals.

It is also unknown how often the system gets it wrong. Meta has not published accuracy figures, how many accounts it has reclassified, or how many adults ended up being treated as teenagers by mistake, what statisticians call a “false positive.” Nor has it explained how to appeal an incorrect classification or how long it keeps the data. Captain Compliance says the agreement with the attorneys general sets maximum error thresholds for these systems, but those figures could not be verified against the text of the agreement.

According to the same consultancy, anyone who wants to change their date of birth from minor to adult must prove it with an ID document or with facial age estimation from Yoti, a company that estimates age from a selfie. This is secondhand information. Yoti's figures help show how difficult this task is: the company itself reported in July 2025 an average error of 1.1 years for young people ages 13 to 17, and says that in a measurement by NIST, the U.S. standards institute, the error was 2.06 years for the 13-to-16 range. In other words, telling whether someone is 15 or 16 — precisely the line that determines what permissions they have — remains an open scientific problem.

Latin America, with no date on the calendar

Since this is a global rollout, the feature should reach the entire region, but none of the sources consulted give dates or a country-by-country timeline. Brazil is the only Latin American precedent mentioned. Nor has Meta announced outside the United States the explainer video that parents there are starting to see in their feed this month, although a guide on the subject published by the organization ConnectSafely can be accessed from anywhere. So far, there have been no known statements from data protection authorities such as Brazil's ANPD, Colombia's SIC or Argentina's AAIP, nor from digital rights or children's organizations in the region.

What comes next is whether Meta publishes data on the system's accuracy and a clear path for correcting errors, and whether the tool reaches Threads, as the company plans before the end of the year. For Latin American families, the conversation about what data on their children an AI analyzes is only just beginning, and it is worth following closely.

Related news

Cybersecurity & Privacy

An official Microsoft package stole credentials on PyPI: what to check if you use AI agents

Two versions of Microsoft's agent-framework reached PyPI carrying a credential stealer, one day after Google warned of attackers using AI to exploit flaws without human intervention. Here is what development and IT teams in Latin America should check right now.

Cybersecurity & Privacy

Microsoft and FAA outages, FBI and USDA breaches: no evidence of a coordinated attack

Over nine days, service outages and alleged intrusions coincided in the United States, and social media buzzed with talk of a “cyber pandemic.” The evidence points to separate technical failures. The proven danger for users is something else: fake emails that use the outages as bait.

Cybersecurity & Privacy

Nginx: two critical HTTP/2 flaws expose millions of servers; here's how to check and protect yours

F5 has fixed two vulnerabilities in Nginx that allow an attacker to take control of a server without a username or password. Public proof-of-concept code is already circulating. This guide explains how to tell whether your server is at risk, which version to update to and what to do while you wait for your system's patch.