Introduction
Recent statistics from the Federal Bureau of Investigation (FBI) reveal that scams have become a pressing issue within the cryptocurrency market, with U.S. citizens suffering losses amounting to $9.3 billion in crypto-related fraud during the past year. This alarming trend has been exacerbated by the developments in artificial intelligence (AI), with blockchain analytics firm TRM Labs reporting a staggering 456% surge in AI-driven scams in 2024 compared to prior years.
The Evolution of Cybercrime
The advancements in generative AI (GenAI) have empowered cybercriminals to execute increasingly sophisticated methods for perpetrating fraud. Cybercriminals now utilize advanced chatbots, deepfake technology, and the ability to create automated token scam networks, thus redefining the landscape of crypto fraud from human-operated schemes to ones driven by complex algorithms that are adaptive and swift.
Ari Redbord, who oversees global policy and government affairs at TRM Labs, explained to Cryptonews the scale at which these generative models are deployed. He noted an emergence of a criminal environment that is not only smarter but also capable of executing numerous scams at once.
By harnessing AI, these nefarious actors can precisely target individuals based on their language preferences, locations, and online activities. In the context of ransomware, AI assists in selecting potential victims who are likely to comply, drafting ransom notes, and even automating negotiations with victims. Moreover, deepfake voices and visuals are utilized in scams that involve impersonating executives or fabricating family emergencies to extract money from victims. Additionally, on-chain scams have evolved such that AI scripts can transfer funds across countless wallets in a matter of seconds—far quicker than any human could manage.
AI-Driven Defense Mechanisms
In response to the rise of AI-powered fraud, the cryptocurrency sector is increasingly embracing its own AI-driven defense mechanisms. Firms specializing in blockchain analytics, cybersecurity, exchanges, and academic institutions are developing machine learning systems designed to detect and thwart fraudulent activities before victims are exploited.
Redbord highlighted that AI is integrated into every core function of TRM Labs’ blockchain intelligence framework, enabling the analysis of vast amounts of data across multiple blockchain environments. This technology allows TRM Labs to identify unusual patterns and behaviors that might signify illicit actions, offering insights that human investigators might overlook.
A similar approach has been adopted by Sardine, a security organization founded in 2020 amid the rise of significant crypto scams. According to Alex Kushnir, the firm’s commercial development lead, Sardine’s AI fraud detection operates on three foundational layers, focusing on gathering in-depth data from every user interaction on digital platforms, connecting with a broad network of reliable data partners, and relying on consortium data for sharing intelligence about known bad actors.
Kushnir emphasized the use of real-time risk assessment tools to act on emerging threats.
Sardine’s Chief of Staff, Matt Vega, noted the efficiency of AI in identifying malicious trends. For example, once a pattern is recognized, AI can conduct deep analysis to formulate recommendations to halt potential threats almost instantaneously—something that would take humans significantly longer. The company collaborates with major crypto exchanges to monitor and analyze unusual user behavior. In one case, TRM Labs reported encountering a live deepfake during a potential financial grooming scam, ultimately verifying the authenticity of the image with AI-assisted tools. Despite their efforts, such scams have still managed to defraud victims of around $60 million.
Future Challenges and User Vigilance
Kidas, another cybersecurity firm, is similarly leveraging AI technology to detect and prevent fraud, as the prevalence of AI-driven scams continues to grow. According to Kidas’ founder Ron Kerbs, their proprietary models focus on real-time analysis of content and inconsistencies, enabling instantaneous risk evaluations to counteract automated scam operations swiftly. Kerbs recounted that just recently, Kidas successfully intercepted two crypto scam attempts on the Discord platform, demonstrating the importance of timely behavioral analytics for safeguarding user accounts.
While these AI detection methods are becoming more sophisticated, experts predict that the threat of AI-powered scams will persist and potentially increase. Kerbs warned that the lowered barriers to sophisticated crimes could allow AI-enabled fraud to become more personalized and widespread as malicious agents may soon orchestrate entire campaigns with minimal human supervision.
In light of these developments, individuals must remain vigilant in protecting themselves against potential scams. Vega advised users to be wary of spoofed websites, citing an incident where a fake website designed to mimic Apple employed Greek alphabet letters to deceive visitors. Users should be cautious of sponsored links and pay close attention to website addresses.
Conclusion
As the industry grapples with the implications of AI on fraud, firms like Sardine and TRM Labs are working closely with regulatory bodies to devise protective measures that employ AI for fraud prevention.
“Our aim is to equip law enforcement and compliance entities with tools that match the speed and sophistication of criminals, enabling us to move from a reactionary approach to proactive risk management,” Redbord concluded.
