Deepfakes Coming to Crypto: An Interview with Pavel Kalaydin, Head of AI at Sumsub
In an exclusive interview with Pavel Kalaydin, Head of AI at Sumsub, we explored the alarming rise of deepfake technology within the cryptocurrency industry. Kalaydin provided valuable insights from Sumsub’s Identity Fraud Report 2023, highlighting the growing threat posed by deepfakes and their impact on Crypto.
The Surge in Deepfake Fraud
Sumsub’s latest report reveals staggering statistics. Deepfake incidents have surged tenfold globally from 2022 to 2023. Alarmingly, 87% of these cases occur within the crypto industry, underscoring the sector’s vulnerability to this sophisticated form of fraud. Kalaydin noted, “The problem with deepfake technology is that bad actors can use it to impersonate others remotely, creating significant security risks. It’s not a new problem; people have been forging signatures for years. Now, with advanced technology, it’s easier and more convincing.”
Regional Highlights: UAE and Saudi Arabia
The report highlights a dramatic increase in deepfake activities in the Middle East. In the UAE, deepfake incidents surged by an astonishing 2200%, while Saudi Arabia experienced a 700% increase. These figures underscore the urgent need for robust security measures in the region. Notably, 36% of deepfake fraud cases in the Middle East and Africa occurred in the UAE, with 11.5% in Saudi Arabia. Kalaydin emphasized that these trends necessitate increased vigilance and advanced technological defenses.
Q1 2024 Data: The Trend Continues
Recent data from Sumsub indicates that the deepfake trend has continued into 2024. In Q1 alone, Iraq saw a 900% rise in deepfake incidents, followed by Saudi Arabia with 600%, Iran with 350%, Israel with 250%, and Kuwait and Yemen with 200%. The UAE experienced a 14% increase during the same period.
Deepfakes in Crypto
The State of Verification in Crypto 2023 report reveals a 128% increase in deepfakes in the crypto industry compared to 2022. Additionally, 70% of crypto companies reported a rise in deepfake fraud, with 55% noting increased fraud-related losses. Kalaydin highlighted that these statistics indicate the growing sophistication and prevalence of deepfake attacks.
Two main approaches to creating deepfakes were discussed. The first involves collecting extensive data on the target and training a model, which, though resource-intensive, is effective. The second, more concerning approach, is zero-shot face warping, which requires only a source video and a single image of the victim. Kalaydin explained, “This method requires no training and can be done with readily available face-swapping software.”
The implications are severe. Kalaydin cited a recent attack on a major crypto exchange where millions of euros were stolen after fraudsters used deepfake technology to impersonate legitimate users. “Platforms implement transaction monitoring to flag suspicious activities. However, deepfake technology allows fraudsters to bypass these checks,” he said.
Collaboration in the Security Space
Despite the competitive nature of the industry, there have been instances of positive collaboration. For example, Sumsub has open-sourced models for detecting synthetic fraud, reflecting a willingness to contribute to the broader security community. This initiative aims to foster feedback and collaboration with academic institutions and other stakeholders.
However, Kalaydin acknowledged that such collaborative efforts are still relatively rare. He noted that while there is some cooperation in areas like Know Your Customer (KYC) and transaction monitoring, it is not yet sufficient. The challenge lies in balancing the need for collaboration with the risk of providing potential attackers with insights into security measures.
Looking ahead, Kalaydin called for more extensive cooperation among security entities. He stressed the importance of verifying partners and ensuring that collaborative efforts do not inadvertently aid malicious actors. Strengthening industry-wide cooperation could lead to more robust defenses against the growing threat of deepfakes.
The Way Forward
Kalaydin emphasized the importance of implementing advanced AI-driven solutions to detect and prevent deepfake fraud. He advocated for the continuous development of sophisticated verification technologies and urged crypto companies to invest in robust security frameworks. As deepfake technology evolves, staying ahead of fraudsters requires constant innovation and collaboration across the industry.
