New research from data and analytics firm LexisNexis Risk Solutions has revealed that over half of UK adults (58 per cent) say they don’t trust mobile apps to accurately identify someone aged 16 to 21 when buying restricted goods or accessing restricted services online.
OFCOM introduces rules this month as part of UK’s Online Safety Act, requiring websites containing adult content to implement strong identity checks to protect young people from harm. A register of categorised services due to be published this summer is likely to prompt requirements for checks on other types of websites in 2026.
Almost half (47 per cent) of respondents raised concern that fraudsters are using AI and deepfake tools to create fake identity documents to fool apps into verifying a user as being over the age of 18.
Even younger age groups showed scepticism, with just a third of 16-24-year-olds (36 per cent) and 25-34-year-olds (34 per cent) saying they are confident in the technology’s accuracy. Another prominent concern amongst these groups is that over 18s could be denied rightful access to services because the technology mistakes them as being too young.
Only six per cent of people aged 55 and over expressed confidence in the technology’s accuracy compared to one in five (19 per cent) of adults overall.
The findings suggest online sales of restricted goods and services could be severely affected where products or services require compulsory identity checks, with 52 per cent of respondents saying they would abandon a purchase online if they had to do a selfie or liveness test and only 28 per cent saying they would not be put off.
Ensuring compliance with the Online Safety Act
Rob Woods, fraud and identity expert for LexisNexis Risk Solutions, commented, “Deepfake-generated identity documents and images can be very convincing, but apps powered by the latest AI tools are well-trained to spot the often-minuscule signs that something is not right.
“The most robust online identity checks don’t just rely on an ID and selfie – they draw information from a wide range of sources to check both traditional identity attributes, like name, physical address and date of birth, as well as digital attributes like your email address and mobile device, which help establish if a presenting identity is genuine or not. Some apps also use powerful tools like proxy piercing to ensure that the identity is not masking their real location, which will be vital for compliance with the Online Safety Act.”
The technology behind identity verification apps is now incredibly sophisticated, with AI-driven tools capable of checking for ‘signs of humanity’ by looking for blood flow under the skin and scanning literally hundreds of features of an identity document in seconds to check its authenticity.
Woods continued: “While natural and understandable with any emerging technology, the public’s perception of these age verification apps is at odds with their reported accuracy levels, which are typically very high at an industry average of almost 97 per cent accuracy. Many fear the new rules will cause a drop-off in customer engagement, but when implemented well, a privacy-first verification step can give customers a positive experience with minimal disruption.
“Like most new technologies, user confidence should increase the more that people get used to the new process. Provided websites are deploying strong AI-supported checks, they needn’t worry about accuracy.”
