Biometric perspectives: Digital identity under real-world scrutiny

November 16, 2026

At a French airport, a digital ID may be enough to check in a bag and board a flight. At the Schengen external border, airport operators and airlines report long waits for biometric checks. These contrasting experiences point to a question facing banks, public authorities and businesses as well: when does identity verification simplify a process, and when does it merely add another checkpoint?
The app may suffice at the gate. Border rules are different.
France allows travellers to use the digital identity card in the France Identité app when checking in baggage and boarding at French airports. This verifies identity for the airline. Where entry rules require a physical travel document, that document is still needed. The app can replace the presentation of a document in a specific process; it does not replace border control or automatically introduce biometric matching at the gate.
The EU Entry/Exit System (EES) serves a different purpose. It records the entry and exit of non-EU nationals making short stays and also processes biometric data. In October, industry bodies ACI Europe and Airlines for Europe reported regular waiting times of more than two hours, with peaks of up to five hours. These are figures reported by the associations, not averages for all border crossings. In November, Biometric Update also reported a legal dispute with the system’s original contractors and the transfer of the central system to a new consortium. The dispute concerns the central IT system, not the separately contracted biometric matching system. The delays therefore cannot be attributed to biometrics alone.

Germany and Latvia ask different questions about facial data

In October, the German Bundestag passed the Digital Identities Act. According to Biometric Update’s report, the approved version allows users to decide whether to include a facial image in the German wallet. Latvia is addressing a different issue: it notified the Smart-ID identification system under eIDAS at the high assurance level. For remote registration, the provider combines verification of the identity document’s chip with facial matching and liveness detection. Users can therefore prove their identity remotely without attending an in-person verification centre.
The examples concern two distinct decisions. First, how should a provider establish identity reliably at registration? Second, which data should later be available in a wallet or presented to a service? Facial matching during registration does not mean that an employer, retailer or public authority must see or check the user’s face every time the identity is used. In Italy, the data protection authority has approved the technical plan for the national IT wallet. Its opinion also highlights the importance of defining responsibility for data-subject rights and protecting sensitive information.

A qualification travels more easily than a professional licence

The European Commission’s proposed Skills Portability Act would allow educational and professional qualifications to be issued as free digital credentials in the European identity wallet. Employers and public authorities could verify their authenticity and validity across borders. The proposal also provides for revocations to be communicated across the EU within 24 hours. A digital qualification does not automatically grant the right to practise a regulated profession: it simplifies verification, while the decision on professional recognition remains separate.
Payments raise a similar question of responsibility. The European Payments Council is working with the SCARLET initiative on common rules for using European identity wallets for strong customer authentication, including the allocation of liability. Two other cases illustrate how such applications might develop. In the Netherlands, a consortium of banks, an insurer and other organisations has developed the MyCheck wallet. In Bucharest, the WE BUILD consortium demonstrated pilot applications for payments and business credentials. Neither the provider-led project nor the pilots demonstrate widespread adoption. They do show that wallet use cases extend beyond government identity documents.

A facial search is a lead, not an investigative outcome

Italy has regulated police use of AI through a decree that came into force at the end of September. Real-time biometric identification is restricted to exceptional cases, must be time-limited and requires prior authorisation. This differs from the retrospective facial searches examined by the UK police inspectorate. There, more than 300,000 searches a year were set against a finding that limits what the figure can show: authorities were often unable to demonstrate reliably how searches contributed to investigations. Potential matches were not always checked adequately.
The two cases highlight different oversight requirements. Before a real-time search, authorities must establish that it is lawful in the specific circumstances. After a search, they must be able to trace what happened to any match. The same principle applies to data incidents: findings should not be overstated. In the Philippines, the data protection commission confirmed reports involving ten agencies, but did not confirm claims that biometric data had been stolen from the transport ministry. In Illinois, a US federal court allowed a class-action lawsuit against Amazon over facial biometrics in Amazon Photos to proceed. This is not a ruling that Amazon violated the law.

An ID system must reach the people it serves

South Africa exceeded its target for replacing biometric smart ID cards more than four months ahead of schedule. The service is available at 753 bank branches, and more than one million people have used the new access point, according to the Department of Home Affairs. In Kenya and Tanzania, by contrast, hundreds of thousands and around two million already-produced identity cards respectively were awaiting collection. These figures should not be treated as comparable usage rates. They do show why the number of cards issued alone is not enough to assess the success of an identity programme.
Burkina Faso began biometric registration in November for a lifelong identification number. Mobile teams are intended to reach people who cannot easily visit a fixed registration centre. Unlike the South African programme, this is the initial creation of a register, rather than the expansion of collection points for an existing identity document. A third approach is to reuse technical components. The UN International Computing Centre (UNICC) developed a digital proof-of-life service with the UN pension fund and contributed to Malawi’s Nzika identity wallet. Reusing technology and expertise does not mean that the services use the same personal data.
Migration administration illustrates how widely the purposes of these systems can differ. The US Citizenship and Immigration Services (USCIS) plans a procurement estimated at USD 200–250 million to bring together, among other functions, biometric data management, identity matching and appointment scheduling. No contract of this size had yet been awarded. South Africa is meanwhile testing a digital asylum process. Its first phase records an intention to apply alongside personal details and facial biometrics. Registration is not a decision on an asylum claim.

Recognising the right person is not enough to authorise a payment

South Korea’s data protection authority has included passkeys in its guidance. Passwordless sign-in with a passkey does not necessarily require biometric matching. At the same time, South Korean company Hancomwith has demonstrated how risk can be assessed continuously during a mobile transfer. In the United States, Trust Stamp uses a separate live selfie approval for payments initiated by an AI agent. This can check that the account holder is present. If the agent subsequently changes the amount or recipient, however, the altered instruction must be authorised as well.
A transfer confirmed by the account holder may still be the result of deception. Incognia is therefore examining longer-term location patterns for indications of the relationship between sender and recipient. The results reported by the provider come from its own data analysis; they are not an independent assessment of effectiveness in live operations. The case illustrates the limits of identity checks alone: a bank may confirm that the right person is present without knowing whether that person is being manipulated into paying a fraudster.

Age assurance and content authenticity are different tasks

The principle that a service should request only the information it needs is especially relevant to age assurance. If the requirement is simply to confirm that someone meets an age threshold, a verifiable proof of that threshold may be sufficient without disclosing the person’s date of birth. The UK media regulator Ofcom, meanwhile, found no statistically significant change in the overall proportion of children encountering harmful online content during the first nine months after child-safety duties came into force. This was not an isolated assessment of age checks. It shows why the effectiveness of online child protection cannot be measured simply by counting verification checks.
The same care in interpreting figures is needed in debates about AI and election advertising. At least USD 137 million cited ahead of the US midterm elections referred to political advertising created partly with AI. The figure is neither a measure of how widely misleading deepfakes were distributed nor of their effect on voting behaviour. Age assurance establishes who should have access; provenance checks establish where content came from. Neither determines whether a claim is true.

Market growth is uneven

Company results also present a mixed picture. US provider ROC increased quarterly revenue by 50 per cent to USD 8 million and returned to net profit. Intellicheck reported a five per cent decline in revenue but higher net profit. In Sweden, Precise Biometrics’ revenue rose by 23 per cent following its merger with Fingerprint Cards, while its operating loss widened to SEK 32 million. Austriacard reported a six per cent year-on-year increase in quarterly revenue. Over the first nine months, however, EBITDA declined despite revenue growth.
The figures show that business growth and improved financial performance are not the same. Whether pilot projects, new rules and additional checks turn into sustainable business will be determined in day-to-day operations. [ML]

This article is based on reporting published by Biometric Update.

Related Articles

Engineered disorder could make protective materials more resilient

Researchers at MIT and Northeastern University have developed a metamaterial with an irregular network structure that withstands impacts better than conventional ordered lattices. The findings could inform future helmet and body armour designs, although testing has so...

IAA MOBILITY 2027: Early bookings signal strong demand in Munich

Around ten months before the event opens, exhibitor registrations are well ahead of the comparable point in the previous cycle. The 2027 programme will put AI, autonomous driving, battery technology and industrial transformation in focus. By 12 November 2026, 166...

Netherlands: Operation Endgame targets three malware networks

International investigators have taken down infrastructure linked to Gootloader, CastleRAT and StealC. The Dutch police report action against 14 Gootloader servers, six CastleRAT servers and 80 StealC servers. Working with international partners, they also seized...

Share This