Yoti responds to Ofcom’s counterfactual statements on age assurance tech

Comments from the UK’s Office of Communications (Ofcom) regarding age verification display a marked lack of understanding around the science and capacities of existing age verification and age estimation tools.

In comments made on the BBC’s Today programme, Ofcom Chief Executive Melanie Dawes says “age assurance technologies which scan your face and estimate your age don’t work very well on children because children can look so different at different ages.”

Yoti’s Robin Tombs has something to say about that. In comments emailed to Biometric Update, Tombs notes that Yoti has been publishing its facial age performance by year of age for 6-12 year-olds since May 2021. “We can confidently say that with millions of individuals’ age data collected from the Yoti app for 13 year olds upwards, plus other diverse consented images of children under 13, there is strong evidence that checking whether children are over or under 13 can be effective.”

“Our scientific data, which could be independently tested, does not support Ofcom’s view that facial age estimation ‘doesn’t work very well on children’.”

The U.S. The National Institute of Standards and Technology (NIST) recently released a summary assessment of facial age estimation with selfie biometrics at the Global Age Assurance Standards Summit. Yoti was among firms tested in NIST’s Face Analysis Technology Evaluation (FATE). Tests did not include images of users under 14 – but could come into play in the increasingly contested territory between 14 and 16. 

Dawes, in her comments noting the perceived “constraints” of facial age estimation for children, also notes that “kids often don’t have that hard ID backup” required to enroll in biometric age verification systems – a more accurate assessment than her stance on what the technology itself can do.

She also emphasizes that Ofcom has issued its response to criticism of its online safety measures for kids as proposals for consultation, and that the regulator welcomes input from industry stakeholders.

The criticism in question is a letter addressed to Ofcom by eight peers in the House of Lords, including online safety advocate Baroness Beeban Kidron, warning that if the regulator does not use the powers it has to regulate social media and other adult content sites, people will come to see the regulation as toothless and poorly designed. The letter lists biometrics-based facial estimation tools among potential “highly effective” solutions to the problem.

Australian age verification debate shifts focus to social media

Australia’s politicians are also struggling with doubts about age verification technologies, although their concerns hinge on privacy and data protection rather than functionality.

InnovationAus.com reports that Prime Minister Anthony Albanese supports the idea of banning social media for children under 16 – “if it can be effective.” His adversary, Opposition Leader Peter Dutton, has taken a more aggressive policy stance, promising to impose and enforce a ban on social media for users under 16 within his first 100 days in office, should he topple Albanese in the country’s next election in 2025.

Asked how he would accomplish his ban, Dutton says “there’s a lot of technology in place to help with the age verification that’s not going to take people’s data…I think that we have a huge capacity here within our system, the technology is much better than it was five years ago, facial recognition to determine somebody’s age is appropriate.”

Per InnovationAus, currently available options for online age verification “can be grouped into three broad categories: those involving the supply of identity documents directly to tech firms, those involving a third party such as a bank verifying the identity of a user, and those using technology to conduct age estimations.”

However, it says, “none of these are presently without significant drawbacks and issues, with concerns over impact on privacy, effectiveness and the potential to be easily bypassed using VPNs.” The latter, in particular, has been cited as an easy workaround for tech-savvy digital natives who are familiar with the Tor network and other readily available Virtual Private Networks.

The Australian list has significant overlap with options that the UK has explored in its age assurance push, in which the only hard rule governing options is that, whatever age check method firms choose to use, it must be “highly effective.”

A roadmap to online age verification developed by Australia’s eSafety Commissioner last year deemed the age assurance market “immature but developing.” To support further testing, the government has pumped $6.5 million into an age verification pilot program. The funding was spurred by concerns over minors accessing pornography, which is seen to be contributing to an increase in violence against women.

But increasingly, the discussion also encompasses social media.

Some big players have made preemptive moves. This month, Meta deployed Yoti’s age verification for Australian Facebook users attempting to edit their age to gain access. The social network has plans for a global rollout of the age verification feature on Facebook and already deploys it on Instagram.

Spain to increase user age to 16 for creating social media accounts

The move will likely be welcomed by regulators worldwide, many of whom are reconsidering age assurance policies for social media. Sur reports that Spain is looking to raise the minimum legal age to create a social media account from 14 to 16. A draft bill outlining the proposal also includes wide-ranging new rules governing the right to a safe and enjoyable digital environment for minors, digital literacy, digital addictions, digital restraining orders, parental controls and the use of fake or fraudulent digital identities.

Per Sur, the draft bill “foresees that the parental control device forms a protective triangle with the duty of providers to use an effective age verification system to provide access to adult content (pornography, gambling or violent content, among others) and with their also new obligation to label, with clear information and simple language, all inappropriate content for minors as well as to publicize the risks for these young people of their products and functions.”

Elsewhere in Europe, France now requires parental consent for registering social media users under 15.

Article Topics

age estimation  |  age verification  |  children  |  Meta  |  social media  |  Yoti