Face it: deepfakes are better and riskier than we think, says Jumio

Never at a loss for hubris, humans tend to overestimate their ability to detect deepfakes, says new research from Jumio. For the identity verification firm’s 2023 Online Identity Study, they looked at data from 8,055 consumers across the U.S., UK, Singapore and Mexico.

More than two thirds of respondents knew about the capabilities of generative AI tools, and more than half were confident that they could spot a deepfake video created by a large language model, or AI. However, advances in generative AI tools (of which ChatGPT is currently the most publicized and well-known) have complicated the picture, and fraud remains an expensive problem. According to Jumio, in 2022, impersonation scams cost the UK 177 million pounds and U.S. consumers $2.6 billion. Clearly, we aren’t as smart as we think we are when it comes to AI – nor as careful as we should be.

“A lot of people seem to think they can spot a deepfake,” said Stuart Wells, Jumio’s chief technology officer, in a news release. “While there are certainly tell-tale signs to look for, deepfakes are getting exponentially better all the time and are becoming increasingly difficult to detect without the aid of AI.”

“While AI-powered technology will increasingly be required by businesses to spot and protect their networks and customers from deepfakes, consumers can protect themselves by treating provocative images, videos and audio with skepticism,” said Wells.

We are, it seems, catching on to the risk, at least in some places. Respondents in Singapore and Mexico demonstrated high awareness of the potential threat of identity theft and other harmful uses, at 73 and 62 percent, respectively. Both the U.S. and UK answered below 50 percent.

“Organizations have a duty to educate their customers on the nuances of generative AI technologies to help them develop more realistic expectations of their ability to detect deepfakes,” said Philipp Pointner, Jumio’s chief of digital identity. However, he stressed the equally urgent need for adequate security tools and data safeguards. “Online organizations must look to implement multimodal, biometric-based verification systems that can detect deepfakes and prevent stolen personal information from being used,” he said.

In results mirrored in similar reports, 68 percent of consumers said they were comfortable using a digital ID, with financial services, government and healthcare ranking among services for which digital ID is preferred.

In early 2023, Jumio announced it had surpassed $200 million in bookings for its selfie biometrics, identity proofing and risk assessment tools.

Article Topics

biometrics  |  deepfake detection  |  deepfakes  |  fraud prevention  |  generative AI  |  Jumio