
Opening a bank account once meant presenting documents across a desk and speaking to a member of staff. Today, many financial services can be accessed through a phone. A customer may photograph an identity document, take a selfie and wait while software compares the face with the document image. Some services also use facial recognition to confirm a customer’s identity during later transactions.
- Where facial recognition enters banking
- The security argument
- Deepfakes and remote impersonation
- What happens if the system gets it wrong?
- Biometric information is not an ordinary password
- Does a face count as strong authentication?
- Who benefits and who could be left out?
- What responsible biometric banking should look like
- Convenience must not become compulsion
- Sources and further reading
Biometric checks can make remote banking more convenient and help prevent criminals from impersonating someone during account opening. But they also create new risks. A face is personal information that cannot be replaced like a password, and a system that fails to recognise a legitimate customer can become a barrier to accessing essential financial services.
Where facial recognition enters banking
One common use is remote onboarding. When a person applies for an account online, a provider may ask them to scan an identity document and capture a live facial image. The system compares the two and may perform additional checks to establish that the applicant is physically present rather than presenting a photograph or video.
Some banks also allow customers to use a phone’s biometric security to access an app. In these cases, the facial template may be held and matched locally by the device rather than sent to the bank each time. That distinction is important: a phone’s face-unlock system, a bank’s remote identity-verification service and a facial payment network are different systems with different risks and responsibilities.
In some markets, facial biometrics have also been used to authorise payments at a point of sale. Such services may connect a face scan to a payment account or stored credentials. They are not the same as the ordinary contactless card terminal shown in the accompanying image, but they illustrate how biometric identification could become another way to initiate or verify a transaction.
The security argument
Financial institutions have a strong reason to make remote identity checks difficult to manipulate. Criminals may use stolen documents, synthetic identities, account takeovers or impersonation to open accounts and move money. A facial comparison can help connect the person making an application to the identity document they claim to own.
The UK government’s guidance on using digital identities under the Money Laundering Regulations, published in February 2026, explains how certified digital identity services can support customer due diligence. Guidance on identity proofing also describes biometric matching, independent testing, false-match rates and enhanced liveness checks as part of a robust process.
These systems can add useful evidence, but they should not be treated as infallible. A face match does not prove that a person is acting freely, that their device has not been compromised or that the underlying identity document is genuine. Effective fraud prevention combines multiple checks and risk-based review.
Deepfakes and remote impersonation
Remote verification creates a particular challenge because the provider is not physically present with the applicant. A criminal may attempt to use a printed photograph, a replayed video, a manipulated camera feed or a synthetic face. More advanced attacks may target the software or device that captures and transmits the image rather than the matching algorithm itself.
Providers therefore use liveness detection and anti-spoofing techniques. These may ask a customer to move, respond to a random prompt or record a short video. Other systems analyse image features or use additional signals to assess whether the face is genuine. The quality of these controls varies, and organisations should test them against current attack methods rather than rely on marketing claims.
There is a trade-off. More demanding checks may reduce some forms of fraud but can also make the process harder for genuine customers. A person using an older phone, poor lighting or an unstable internet connection may struggle to complete a scan even when their identity is entirely legitimate.
What happens if the system gets it wrong?
A false rejection can prevent someone from opening an account, accessing a service or completing an urgent transaction. A false acceptance can allow an impostor through. Both errors matter, but the consequences are different and depend on the context in which the system is used.
Some people may find facial verification difficult because of disability, facial differences, medical equipment, age-related changes or other circumstances. In June 2026, the UK government told Parliament that certain identity-checking methods will not suit every user and that the digital verification framework allows a range of approaches, including methods that do not require facial biometrics.
The government also said financial services providers must consider their obligations under the Equality Act 2010 and make reasonable adjustments where necessary. That is important because an alternative route should be practical, accessible and available without unfair delay or extra burden.
Biometric information is not an ordinary password
Facial data used to uniquely identify a person is treated as special-category biometric data under UK data-protection law. Organisations need an appropriate lawful basis and a separate condition for processing this particularly sensitive information. They must also explain the purpose, limit collection, secure the data and retain it only as long as necessary.
There are different technical models. Some systems store a mathematical template rather than a photograph. Some keep the template on a personal device, while others may send information to a service provider or maintain central records. A template is not necessarily reversible into a normal photograph, but it remains sensitive data and should not be treated as harmless simply because it is encoded.
Customers should be able to find out which organisation controls the data, whether a third-party provider is involved, where information is held, how long it is retained and what happens if a security breach occurs. Clear answers are especially important when a face scan is required to access an essential service.
Does a face count as strong authentication?
Payment security rules distinguish between different authentication factors. These are commonly described as something a person knows, something they possess and something they are. A biometric characteristic can form an inherence factor, but strong customer authentication generally requires a combination of at least two independent categories where the rules apply.
The UK’s Financial Conduct Authority explains the strong customer authentication framework for payment services. A face scan alone should not be confused with a complete security process. The way the biometric is linked to a device or account, and the other factors used to authorise access or a payment, are part of the overall protection.
For many customers, facial recognition is only one stage of a wider process. A bank may use a phone’s secure hardware, a passcode, a registered device or transaction monitoring alongside biometric authentication. The important question is whether the combined controls actually reduce the risk of unauthorised access.
Who benefits and who could be left out?
Biometric verification can help people who find paperwork difficult or who cannot easily visit a branch. Remote onboarding may be faster and less expensive, and a well-designed system can make services available beyond normal opening hours.
But a digital-only process can exclude people who do not own a compatible smartphone, have limited digital skills, cannot hold a device steadily or cannot complete a facial scan. If a bank removes face-to-face alternatives or makes the biometric route the only practical option, convenience for some customers can become a barrier for others.
Financial inclusion requires more than a technically functioning app. Providers need accessible alternatives, clear support and a way for customers to resolve failed checks without repeatedly submitting sensitive images. Staff should be able to recognise when an automated process is unsuitable and move a customer to another route.
What responsible biometric banking should look like
Financial institutions should explain why a face scan is necessary and whether another method is available. They should test systems for spoofing, false matches and demographic performance, keep biometric data secure and minimise retention. Customers should not be forced to accept unrelated uses of their facial information as a condition of receiving a basic service.
Independent assessment matters. Providers should monitor how often customers fail verification, how many cases require manual intervention and whether particular groups experience disproportionate difficulties. They should also review systems as fraud techniques evolve, rather than assuming that a successful launch guarantees long-term security.
Convenience must not become compulsion
Facial biometrics can strengthen parts of the banking process, particularly where they are combined with reliable documents, device security and effective anti-spoofing checks. They can also reduce friction for customers who prefer to manage their finances remotely.
But the face is not a universal key. Some people cannot use facial recognition reliably, and even sophisticated systems can be attacked or make mistakes. A trustworthy financial service should protect accounts without making access dependent on a single biometric method. The future of biometric banking should be measured by security, privacy and inclusion together.
Sources and further reading
- GOV.UK: Using digital identities with the Money Laundering Regulations (26 February 2026)
- GOV.UK: How to prove and verify someone’s identity
- UK Parliament: Facial recognition verification and barriers for disabled people (answer, 5 June 2026)
- UK Parliament: Alternative identity checks for financial services (answer, 4 June 2026)
- Financial Conduct Authority: Strong Customer Authentication
- Financial Conduct Authority: Comparison of banking providers’ fraud controls
- European Banking Authority: Requirements for inherence in strong customer authentication
