Facial Recognition Technology
AI Cybersecurity Europe

Who’s Who? ID R&D Tests How Well an AI Computer Can Spot Fake Faces

In a recent test of Spot the Spoof by the face biometrics provider ID R&D, out of 175,000 images, a group of AI-powered computers were able to identify the false faces amongst the real ones every time and with 0% failure; far surpassing the ability of any human when put through the same task.  

Facial recognition has become a popular form of personal authentication and is deemed to be highly secure. From unlocking a smartphone to making a payment, to finding missing people, facial recognition’s use cases are familiar far and wide.

On the face of it, this form of software appears to be impenetrable, yet as the technology evolves so too do the techniques that fraudsters use to break in. They very often exploit things like printed photos, videos, digital images and 2D or 3D masks as a means to override the system.

To explore the capabilities of an AI machine alongside those of a real person in spotting the false facades fraudsters frequent, ID R&D conducted a test, the results of which have recently been released in its ‘Human vs Machine: Can people spot spoofs better than AI?’ report.

Machines outperformed humans for all types of spoofing techniques

The test found that a computer consistently outclassed humans across all five techniques by scoring 0% error rates when faced with all 175,000 images, and all types of attack.

The results humans provided weren’t nearly as cut clean. The test showed that they had a far lower degree of accuracy for every type of spoofing technique, including misidentifying 30% of printed photo spoofs.

Even when a group of seventeen humans deliberated over the images, resulting in a more accurate outcome than an individual human, their majority decisions were never better than the computer’s performance on the same task.

The test also found that computers were almost 20 times quicker to determine liveness. On average, the computers running on one CPU core were able to make split-second decisions, taking 0.5 seconds to process each image.

However, even though they were able to achieve this fast feat, when considered that they were hooked up to a central processing unit, at this rate, it would still have taken the computers just over 24 hours to run through all 175,000 images.

When the number of people using facial recognition to secure payments is expected to exceed 1.4 billion by 2025, many will argue that this time is just not fast enough, and I’m sure that speed will continue to be a key focal point for AI developers as we head into 2022.

However, the computers’ results were far superior to those produced by a human. The study found that it took humans an average of 4.8 seconds to process each image.

This performance proves to organisations in financial services and in other industries that they should continue to trust the technology. Their reliance upon it for layers of fraud detection can save time and enable human resources to be focussed on other areas of the consumer lifecycle.

Technology ensures the best frictionless customer experience

Despite the strong performance of computers at spotting spoofs, overzealous fraud detection must not compromise the experience of genuine customers. Many facial recognition systems on the market achieve low rates for letting fraudsters in by making it likely that genuine users are also caught in the net.

However, in this study, the AI machine erroneously classified just 1% of genuine faces as spoofs. Humans, on the other hand, misclassified 18% of genuine faces as spoofs, proving that automation technology is also better than humans at keeping genuine users out of the fraud net.

Organisations may rest assured that technology has become the best way to ensure frictionless verification experiences for customers.

Alexey Khitrov, CEO, ID R&D
Alexey Khitrov, CEO, ID R&D

“The results are undeniable; we can leverage technology to keep us safe from fraudsters,” said Alexey Khitrov, CEO at ID R&D. “Biometric technology has undergone significant evolution in recent years to increase speed and accuracy, outperforming the human eye. Organisations can make fantastic efficiencies by using biometric systems. However, there is still work to be done. The technologies must continue to find the balance between security and convenience, by reducing fraud to the minimum amount possible while ensuring genuine customers have a perfectly smooth experience.”


  • Tyler is a fintech journalist with specific interests in online banking and emerging AI technologies. He began his career writing with a plethora of national and international publications.

Related posts

Greater Manchester has largest regional FinTech ecosystem [Report] & [Map]

Jason Williams

Is Credit Card Usage Experiencing 2008 Déjà Vu Amid Current Crises?

The Fintech Times

UK Fintech News Roundup: The Latest Stories 31/08

Tom Bleach