The rise of artificial intelligence is fueling a sharp increase in crypto-related fraud, with deepfake technology driving a surge in sophisticated scams, according to a new report.
In the first quarter of 2025, at least 87 scam rings using AI-generated deepfakes were dismantled, according to the 2025 Anti-Scam Research Report co-authored by crypto exchange Bitget, blockchain security firm SlowMist, and analytics provider Elliptic.
The report also reveals crypto scams reached $4.6 billion in 2024, marking a 24% increase from the year before. Nearly 40% of high-value fraud cases involved deepfake technologies, with scammers increasingly using sophisticated impersonations of public figures, founders and platform executives to deceive users.
Related: How AI and deepfakes are fueling new cryptocurrency scams
“The speed at which scammers can now generate synthetic videos, coupled with the viral nature of social media, gives deepfakes a unique advantage in both reach and believability,” said Gracy Chen, the CEO of Bitget.
Deepfakes pose new threats
The report details the anatomy of modern crypto scams, pointing to three dominant categories: AI-generated deepfake impersonations, social engineering schemes, and Ponzi-style frauds disguised as decentralized finance (DeFi) or GameFi projects. Deepfakes, it warned, are especially hard to detect.
AI can simulate text, voice messages, facial expressions and even actions. For example, fake video endorsements of investment platforms from public figures such as Singapore’s prime minister and Elon Musk are tactics used to exploit public trust via Telegram, X and other social media platforms.
AI can even simulate real-time reactions, making these scams increasingly difficult to distinguish from reality. Sandeep Nailwal, co-founder of the blockchain platform Polygon, raised the alarm in a May 13 post on X, revealing that bad actors had been impersonating him via Zoom.
He mentioned that several people had contacted him on Telegram asking if he was on a Zoom call with them and whether he was requesting them to install a script.
Related: AI scammers are now impersonating US government bigwigs, says FBI
SlowMist CEO Yu Xian urged users to verify the legitimacy of Zoom links and domain names to avoid falling victim to video-based impersonation scams.
New scam threats call for smarter defenses
As deepfake scams grow more convincing, security experts say education and vigilance are critical. For institutions, regular security training and strong technical defenses are essential, the report stated. Businesses are advised to run phishing simulations, protect email systems and monitor code for leaks. Building a security-first culture — where employees verify before they trust — is the best way to stop scams before they start, the report added.
Chen offered everyday users a straightforward approach: “Verify, isolate, and slow down.” She added:
“Always verify information through official websites or trusted social media accounts — never rely on links shared in Telegram chats or Twitter comments.”
Chen also stressed the importance of isolating risky actions by using separate wallets when exploring new platforms.
Magazine: Baby boomers worth $79T are finally getting on board with Bitcoin
