Age estimation tech faces an uphill battle in Australia

Last week, the Australian government announced an age assurance trial in preparation for introducing rules for regulating children’s access to age-restricted content. Age assurance technology vendors are now facing an uphill battle in convincing not only the government but also young social media users and their parents that their products work.

One of the main question that the AU$6.5 million (US$4.3 million) Age Assurance Technology Trial will need to answer is whether technologies determining whether someone is 18, 16, or 13 years old are effective enough – including biometric age estimation.

Regulators in countries such as the UK have cast doubt on whether age estimation tech is accurate enough to be deployed as part of its Online Safety Act. Earlier this year, the country’s communications regulator Ofcom noted that the age estimation technology was accurate at determining who is under or over 18, but not “between children in different age groups to a highly effective standard,” The Times reports.

However, industry players such as the Age Verification Provider’s Association (AVPA) and age verification company Yoti are pushing back against the claim that age estimation needs to be “highly effective” to be efficient.

“We strongly disagree with Ofcom that facial age estimation (FAE) does not work very well for children, a claim they have not substantiated with any scientific evidence,” says Robin Tombs, Yoti co-founder and CEO, in a LinkedIn post published last week.

The key question for Tombs is: “How good is good enough?” The industry already has ways to provide highly effective age estimation by scanning ID documents with attack detection (PAD) and injection attack detection (IAD). But trying to secure perfection is often the enemy of effective regulation, he notes.

“For social media users around the age of 13 regulators should use ‘broadly effective’ age checks,” says Tombs.

Whether this argument will convince Australian regulators remains to be seen. Some signs point towards a positive outcome for facial estimation tech in the country.

Earlier this month, eSafety commissioner Julie Inman Grant and Robert French, the former chief justice of the High Court of Australia published a legal examination on banning children’s access to social media through the planned Online Safety Act 2021.

“The technologies for age assurance are evolving,” the report says. But it also adds that facial age estimation is the “most viable and privacy-preserving within the biometrics category.”

Aside from recommending the age assurance trial, the legal examination also proposes that a regulatory framework should be established for accreditation and oversight of age assurance providers.

Meanwhile, Michelle Rowland, Australia’s minister for communications has noted that using official documents such as a birth certificate or the government digital ID system may not be appropriate for checking the age of social media users due to privacy and security concerns.

The age assurance trials will be conducted by testing laboratories designated by the Australian government with results expected in Spring of 2025. The results are intended to inform the government’s review of the Online Safety Act. Aside from social platforms, the administration is also looking at other use cases of age-restricted content, such as pornography and video games.

Last year, the Albanese government in Australia decided against implementing a mandatory age verification system for online pornography and other adult content, citing the underdeveloped state of existing technology solutions.

Questions on social media age bans still unsolved

While age estimation industry players are arguing whether social media should use highly effective” or “broadly effective” age assurance methods, Australian regulators still disagree on which ages should be restricted at all.

“There’s no clear guidance anywhere in the world about what is an appropriate age,” Rowland, Australia’s minister for communications said in an interview for ABC last week.

Robert French, former High Court Chief Justice has been rooting for banning children under 13, while those of 14 and 15 would be subject to parental consent. Prime Minister Anthony Albanese has floated a ban for kids up to age 16. The decision, however, will need to be supported across the Parliament, says Rowland.

The Minister also said that the technical trial for age assurance between the ages 13 to 16 will look at its efficiency for ethnicity. But the trial is not just testing efficiency, it is also measuring the acceptance of the technology.

“This is important to see the acceptance by the platforms, the acceptance by the public, to actually use them and how effective they really are,” says Rowland.

According to a survey published by a group of Australian scholars at the University of Sidney, introducing age restrictions on social media has invited mixed reactions from young people and their parents.

According to the research, 72 percent of teenagers aged 12 to 17 and 86 percent of parents say more effective age limits would improve online safety for young people. The survey, however, also revealed some drawbacks, The Conversation reports.

“Our findings suggest age verification is generally supported, but participants think it likely would not work,” the researchers say. “Instead, they said more safety education, face-to-face dialogue, and accountability from social media companies would be better approaches to keeping young people safe online.”

Article Topics

age estimation  |  age verification  |  Australia  |  biometrics  |  children  |  consumer adoption  |  face biometrics  |  facial analysis  |  social media  |  Yoti