Fears of fraud are deepening across all verticals as digital onboarding and passwordless authentication become more popular. These fears are spurring interest in biometric liveness detection, which aids in the identity verification process. But providers of this technology are still working to differentiate and perfect their solutions.
In a non-digital environment, personal interaction minimizes fraud—for example, a live customer walking into a bank branch would likely struggle to open up accounts with five separate driver’s licenses, all bearing different photos. Digital tools, for all their convenience, are making it easier for fraudsters to spoof identities and automate fraudulent transactions.
That increase in potential fraud is driving colossal market demand for biometrics, particularly facial recognition technologies. These technologies confirm a user’s identity based on facial features—features unique to each user, which the user can conveniently demonstrate using a phone or laptop camera. But facial recognition technology itself can be spoofed; for example, a bad actor can simply steal a high-res image of a consumer along with their other personal data and use that image to pass facial recognition tests.
This is where liveness detection enters the discussion. Liveness detection is the ability to detect the users’ live presence, determining if it is a real person at the point of capture instead of fake representations. The technology can include capabilities like motion detection, multi-frame image processing, 3-D image mapping, and more. Liveness detection technology can work alongside, or separately from, the identity verification process, providing an extra layer of security and preventing access to bad actors.
Our Outside in Report: Liveness Detection takes stock of the market for liveness detection technologies, including current and future use cases and barriers to the technologies' adaptation. Our report shows that liveness detection is a growing market, as the technology appears to be the next step in the evolution of identity verification. Still, apparent challenges prevent the technology from being ubiquitous, including process friction and competition from multiple look-alike vendors.
Our research found that, in 2021, the market TAM for biometrics generally was $22 billion, and that number is rising at about 13% CAGR. Most of the market penetration has been in industries with high customer onboarding requirements that understandably benefit from additional fraud protections—think governments and financial services firms.
For example, the banking industry is fuelling growth, especially when it comes to Finserv and the payments industry. In the report, we project a 33% CAGR in biometric solutions through 2025 across the online banking industry.
In government, the rise of eIDs and biometric passport initiatives are leading to widespread adoption, as these tend to require high verification standards. The trend is visible in both developed and developing areas—for example, 22 of 54 African nations currently offer biometric passports.
As biometric solutions grow, so will liveness detection…if the industry can keep up.
Even as the liveness detection market grows, players in the market have noted that there are challenges ahead. And like any challenges, these can represent opportunities to serve the market better and become a best-in-class leader.
Bias. One of the most vocalized issues is that of bias. There have long been concerns over biases within AI-operated biometric technologies, especially if the technologies do not have access to a broad enough set of training data. Biases and lack of data can lead to accuracy gaps related to age, gender, skin tone, and more. For example, an article in Wired magazine from 2019 reported on most AI’s struggles with accurately identifying black faces, leading to a much higher false match rate.
Accuracy. But bias might be as much a symptom as a cause: Overcoming bias is a critical way to improve accuracy. Liveness detection technologies are still relatively new, and vendors are racing to tune their algorithms to provide the most accurate results—while simultaneously bringing down the time to result.
Adam Perold, CEO of Element (a modern AI company focused on digital identity), put it succinctly: “Right now, most solutions in the market are not excellent, and the best stuff is just okay.”
Friction. Bias and accuracy are not the only issues. Identification processes—whether digital or “old school”—have always added an extra “hassle factor,” and consumer patience is slim. “Consumers want processes to be as smooth as possible,” says Bill Spruill, CEO of GDC, “so any small amount of friction can lead to a measurable drop-off – not just from bad actors, but the good ones.” The friction caused by active liveness checks can reach a tipping point if they stall digital processes more than they enable them. This, Spruill says, will diminish the technology's overall value.
It’s no wonder, then, that there is a sort of arms race going on to develop the best-in-breed solution, with no clear winner yet. As Perold notes, “The first 70% [of building a liveness detection solution] is easy, but the difficulty curve really steepens from there.”
For a complete overview of the liveness detection market, including the most mature use cases and players in various verticals, you can download our Outside-in report on Liveness Detection.
Interact with the content you’re truly interested in. Our members receive exclusive and early access to information around market drivers, legislation, investment trends, and consumer insights across the entire data ecosystem. Learn how to receive exclusive access to key market drivers so you’re prepared and informed.