Home Office provides answers on using digital ID to prove age for alcohol sales

On the surface, the UK’s quest to allow adults to use digital ID to prove their age for alcohol sales is a fairly simple proposition: instead of handing the barkeep or shop clerk a plastic card that proves who you are, you can call up a credential on your phone that says your age has been verified, and that you’re OK to buy booze – without providing any other personal information, like an address or biometric.

But this switcheroo, says Tony Allen of the Age Check Certification Scheme (ACCS), has revealed itself to be a much more complicated affair. In a new post on LinkedIn, Allen notes that “the publication of the draft Licensing Act 2003 (Mandatory Licensing Conditions) (Amendment) Order 2026 has generated a much more interesting debate than might have been expected from what, on first reading, looks like a relatively narrow change.”

Allen chairs the Expert Panel on Age Restrictions, which exists to help the government make age restriction regulation work in practice. He says the Panel’s recent review of the draft order aimed to attain clarity on certain aspects – specifically, semantic interpretations, the practical customer journey, self-checkout operation, the meaning of “deliver identification”, data minimization and the assumptions contained in the Economic Note, as well as the broader implications for use of digital ID at self-checkout.

Allen also says a quick response from the Home Office provided most of the clarity the Panel had sought.

No doubt needed for digital ID age assurance transaction

One major point of clarification involves who the rules allow to use digital ID, and in what circumstances. Specifics in the draft language read to the Panel as though the digital ID option could be limited to those whose age is in question. However, the government response has clarified that “the potential use of digital proof of age is not, in the Home Office’s view, limited to the Challenge 25 population.” Retailers can, if they want, “operate a self-checkout process under which every customer purchasing alcohol provides digital proof of age, irrespective of whether anybody has reason to believe that particular customer is under 25.”

Another concerns sections of the draft that seem to circle back on themselves. Based on the Home Office’s response, Allen broadly concludes that digital ID systems deployed at retail should be capable of meeting the statutory requirements for obligatory proof of age, triggered by reason to believe a customer is not old enough to buy alcohol. In other words, systems should have substantive safeguards, including an agreement with a registered Digital Verification Service (DVS) provider, confirmation that the customer has attained the relevant age, the medium confidence requirement and registration under the DVS trust framework.

For legal purpose, ‘identification’ means age assurance

Both age verification algorithms and age assurance laws boil down to language, but the latter are more burdened with the weight of etymology and usage. Human language is slippery, and words that carry massive semantic weight and broad application may mean something very precise and specific in the context of privacy law.

In the context of the draft order, one such word is “identification.” The language now says that a DVS provider must “deliver identification” at a minimum medium level of confidence. That’s problematic for the age assurance industry, which has been working to differentiate age assurance (which just wants to know that a customer is old enough) from identity verification (which needs to verify a person is who they say they are).

To that end, Allen writes that “the Panel queried whether the intended concept was really an age verification result. That question matters because one of the principal attractions of digital proof of age is selective disclosure.” The Panel suggested replacing “identification” with “age verification result” – but the Home Office says it’s not necessary, and that, in keeping with established legal language, “identification” should be understood in this context as “identification for a specific purpose: establishing whether the customer has attained the relevant age.”

“Importantly,” Allen says, “the Home Office confirms that the wording is not intended to require unnecessary personal information to be disclosed.” So, while identification may refer to age assurance, the reverse is not applicable: age assurance is not identity verification.

Self-checkout options grow in light of clarification

The scenario as it stands presents retailers with a choice for self-checkout. Those who do not wish to install facial age estimation tools can opt to require ID checks for every self-checkout purchase. This removes the need for human verification (the clerk comes over and swipes his card to approve), but adds an age check for every customer, rather than just those who look suspiciously young.

“Most customers do not produce any proof of age,” Allen writes. “Universal digital checking means that every customer does. The 22-year-old, the 50-year-old and the 75-year-old all generate a digital transaction. That increases cost and creates a new customer journey for people who today simply receive a visual clearance.”

Allen suggests that this is a “coherent model” that explains why “the Economic Note anticipates very high DVS transaction volumes and substantial reductions in staff intervention.” Technically, the changes mean a store could sell age-restricted products without needing any human cashiers at all.

Age estimation, of course, is also available. Allen notes three practical models accessible to retailers. “The first is traditional Challenge 25 with a human making the initial visual assessment and digital proof replacing physical ID when a challenge occurs. The second is universal digital proof of age: everyone purchasing alcohol through a particular channel proves age digitally.”

The third, which he suggests may offer the greatest combination of automation, proportionality and reduced transaction cost, is “a hybrid model in which age estimation filters out customers clearly above the challenge threshold, while customers closer to that threshold provide stronger proof.”

The problem is, at this time, the government’s published position, which says it does not intend “at this time” to allow age estimation technology to form part of the alcohol age verification process. Allen says that decision was made partly based on the lack of adequate governance structures – which has been remedied by the publication of ISO/IEC 27566-1, the international standard on age assurance technologies.

Overall, Allen says, the standards and governance looks much different than it did when the decision on age estimation was made.

“That does not automatically mean age estimation should now be permitted for alcohol sales. It does mean that the original reason for caution may deserve to be revisited as the surrounding assurance ecosystem develops.”

UK pubs to accept digital ID for age assurance by Christmas 2026?

Allen has been following the government’s progress toward the revised Mandatory Licensing Conditions for a long time. In previous posts, he has called their arrival “a genuinely significant milestone for age assurance in the UK,” in being “the first time that digital identity has been explicitly recognized within one of the UK’s most widely used age-restricted sales regimes.”

But the finish line is still ahead, and obstacles remain – many in the form of ambiguity. “I’ve had conversations with retailers, licensing practitioners and technology specialists,” Allen says, in a post that outlines what would become the Panel’s formal concerns. “And, as often happens with legislation, the more people look at it, the more possible interpretations emerge.”

He also notes that, for a rule that “will need to be understood by tens of thousands of businesses, licensing authorities, Trading Standards officers, police forces, technology providers and eventually courts,” clarity is of utmost importance. Legislation is not designed as a communication tool, but a reference for societal rules and judgment. It prioritizes legal precision over general intelligibility. The complexity of the digital ID issue may warrant that. But in the end, people cannot follow a law they do not understand.

In Allen’s words, “the challenge now is therefore not necessarily to amend the policy. It is to ensure that the final legislation and guidance communicate the intended operating model with enough clarity that everyone reaches the same interpretation.”

Can the government actually get everyone on the same page in the next four and half months, in order to finally deliver on its promise of being able to use digital proof of age for pints at the Musty Badger this Yuletide? The clock behind the bar is ticking, but only the Spirit of Christmas Yet to Come knows when it will finally sound the bell formally welcoming digital ID into the mix.

Article Topics

Age Check Certification Scheme (ACCS)  |  biometric age estimation  |  digital verification service (DVS)  |  facial age estimation (FAE)  |  retail biometrics  |  UK age verification  |  UK digital ID