Feature Alternative Lending

Deepfakes, AI and emerging risks for SME lenders

Emerging technology is sure to find a welcome home in the arms of fraudsters, but when are the risks set to get serious for SME lenders, and is their success inevitable?

a man holding a white card


This is an excerpt from AltFi Research and Equifax's ID Fraud And AML In 2023: An SME Lender’s Guide, which is available for free here.

Compared to 20 years ago, artificial intelligence (AI) and smartphone video recording have made it dramatically easier for many smaller lenders to finalise applications remotely. 

Assurance processes that could once have taken weeks can now be finished in just mere hours. However, it would be naive to think that fraudsters won’t look to take advantage of these same innovations. 

As the underlying technology gets cheaper and more easily accessible, deepfakes—where existing material is used to create a realistic video of someone's likeness—are poised to become a common problem that SME lenders will have to learn how to deal with, just like a forged signature. In addition, machine learning will likely create new ways to forge documents and bypass checks that have become largely automated. 

How Soon Could Issues Emerge?

Deepfakes don’t yet represent “a clear and present danger” to SMEs according to Iain Armstrong, global regulatory affairs practice lead at ComplyAdvantage—but he feels that this could easily change in the future. Armstrong believes the next decade is a good estimate for when these risks will become a continual concern for lenders.

“I think it's a risk that is likely to progressively grow over the course of the next 10 years,” he explains. Armstrong believes the threat posed by deepfakes will be particularly significant for fintechs and challenger lenders, who don’t have any physical or branch presence where they can direct high-risk borrowers to “physically sign something or talk to someone”. 

Looking back 20 years ago, Armstrong points out that, in the past, the real world was where much of the assurance processes took place. “All of that now has to be done through algorithms,” he explains, which presents issues of its own when it comes to verification.

One of the ways deepfakes could impact SMEs, according to Armstrong, is if criminals use these to pass what’s called “liveness” checks, the process where the lender verifies the target is human. In the modern era, these can take the form of uploading videos of the applicant to prove identities, like when you open a Monzo or Binance account.

One area in which deepfakes could see widespread use is in what is known as “ghost fraud”, where financial details of the deceased are used to take out loans, with a deepfake used to complete the application process.

What Do Lenders Need To Be Aware Of?

Dimitrie Dorgan, global head of fraud at SME lender Stenn, wants to highlight how the use of machine learning could make the threat of document forgery even more dangerous, claiming this issue will escalate in the future as the computational power which makes these systems work gets cheaper and tools like ChatGPT only become more accessible and advanced...

Want to keep going? Read the full feature in AltFi Research and Equifax's ID Fraud And AML In 2023: An SME Lender’s Guide, out now!

Companies In This Article

logo, company name

More Like This