The Rise of AI Selfie Age Verification: How Technology Is Transforming Online Access Regulations

Age verification through selfies has become remarkably straightforward for users.(Representational)

United Kingdom:

As governmental regulations tighten on online platforms ranging from social networks to adult content sites, a specific industry is experiencing significant growth: AI-powered age verification using selfies.

This technology offers rapid and precise implementation of legislation such as Australia's prohibition of social media for those under 16, which becomes effective December 10.

From a user perspective, age verification via this technology is exceptionally simple.

The process only requires capturing a front-facing selfie with a smartphone or webcam, after which an automated decision is delivered within 60 seconds.

"We estimated your age is 18 or older," displayed a notification window on the Roblox gaming platform.

At the expansive London headquarters of startup Yoti, mannequin heads—some adorned with wigs or masks—are arranged along windowsills for testing purposes.

Their AI system isn't deceived by these props. "We can't be sure that the image was of a real face," reported one test platform installed on a mobile device.

Over time, "the algorithm got very good at looking at patterns and working out, 'this face with these patterns looks like a 17-year-old or a 28-year-old'... it just got very good at estimating age," explained CEO Robin Tombs.

Data Privacy Concerns

Currently, the startup conducts approximately one million age verifications daily, serving major clients including Meta, TikTok, Sony and Pinterest, according to Tombs.

Yoti achieved profitability this year after recording revenue of 20 million pounds ($26 million) in the 12-month period ending March.

The company projects a 50-percent revenue increase in its current financial year.

Competitors are likewise thriving, with 34 companies such as Persona, Kids Web Services, K-id or VerifyMy counted among members of the Age Verification Providers Association (AVPA).

In 2021, the AVPA projected annual revenue approaching $10 billion for the sector across the 37 OECD member nations during 2031-36, though no more recent forecasts have been issued.

AVPA director Iain Corby remains cautious about making specific predictions.

"There's so much that can happen. Regulation is moving very fast, technology is moving very fast," he told AFP.

AI-driven age verification methods have prompted questions regarding potential bias and privacy implications.

These tools are "potentially a bit intrusive or dangerous to people's private life," noted Olivier Blazy, cybersecurity professor at Polytechnique, the elite French engineering institution.

However, much depends on what information websites share with their external age verification providers, he added.

"I think the ecosystem is tilted towards AI at the moment, but there may be a pull-back in the coming years to better protect users' private life," Blazy suggested.

Make-Up Challenges

Blazy highlighted that these systems still have vulnerabilities, noting that "just with normal makeup, it's relatively easy to make yourself look older or younger".

Others have identified biases in age-verification algorithms, which can be less precise when analyzing non-white faces.

A report from an independent Australian organization found that "underrepresentation of Indigenous populations remains a challenge"—albeit one "that vendors are beginning to address".

Yoti's CEO Robin Tombs acknowledges that less data was available to train their model for certain age groups or skin tones.

However, he maintains their tool can "check to see whether we think people are using false accessories or they are using makeup and other such things".

Regarding privacy, he stated that Yoti's system immediately deletes all data after completing its analysis.

Platforms implementing selfie analysis tools from companies like Yoti can customize settings for evaluations, allowing for error margins.

For instance, websites prohibiting under-18s often require the system to determine that visitors are over 21.

Users falling within ambiguous ranges may be directed to utilize more traditional verification methods—such as photographing or scanning an official identification document.