Traditionally, scamming required a significant amount of manual labor. Whether it was maintaining a fake identity for a romance scam, answering questions as a fake investment advisor, or impersonating someone to trick victims into sending money, human involvement was crucial. However, AI is changing the game by enabling scammers to automate these tasks and scale their operations without the need for additional manpower.

According to FATF president Giles Thomson, AI allows a small team of scammers with access to a powerful server to carry out operations that previously required a much larger workforce. This increased efficiency translates to higher profits for criminals, as they can target more individuals simultaneously while reducing their operating costs.

Chainalysis research has shown that scam operations linked to AI vendors generate significantly higher revenue compared to those without such connections. This highlights the financial incentive for scammers to invest in AI tools that can streamline their fraudulent activities.

AI technology enables scammers to create and maintain fake identities, engage in conversations with multiple targets across different languages, and generate convincing documents and images to support their scams. This automation makes it easier for a single operator to manage multiple conversations and fraud attempts simultaneously, presenting the facade of a larger and more sophisticated operation.

The impact of AI-driven scams is already being felt, with billions of dollars in losses reported due to AI-related fraud schemes. Criminals are leveraging AI models to streamline their operations, as seen in cases where AI tools were used to assist in technical tasks and analyze information for extortion campaigns.

While AI is making it easier for scammers to deceive victims, there are efforts to combat automated fraud. Companies like O2 have developed AI bots to engage with scammers and waste their time, disrupting their operations. However, as scammers also employ AI-driven bots, the effectiveness of such defensive measures may diminish over time.

As the battle between scammers and defenders evolves, the focus is shifting towards identifying and targeting the key components of fraudulent operations that cannot be easily automated. This includes disrupting the financial infrastructure that enables fraudsters to profit from their schemes and implementing measures to prevent unauthorized transfers of funds.

Ultimately, the rise of AI in scam operations poses a significant challenge for both individuals and authorities tasked with combating fraud. As technology continues to advance, it is essential to stay vigilant and employ effective strategies to protect against increasingly sophisticated and automated scams.

TRM Labs’ AI-in-Crime Adoption Index classifies scams as the only category of crypto crime where artificial intelligence has reached a “Mature” level of adoption. The use of AI on the scammer-side, including deepfakes, chatbots, and AI-powered lures, has increased approximately 13-fold since 2022.

This shift exposes a vulnerability that traditional smart-contract security does not address. Even if an exchange account is properly authenticated, a hardware wallet signs correctly, or a smart contract executes as programmed, funds can still end up in the hands of an attacker if a deepfake convinces the user to approve a transaction.

This places more emphasis on security measures before authorization, such as verifying account-recovery requests, approving transfers, or accepting payment instructions from familiar sources. The moment before authorization becomes crucial in preventing deepfake scams.

TRM’s index measures the prevalence of AI in different types of crime, the stages at which it is used, and the sophistication of the tools involved. The use of AI in scams has increased significantly since 2022, with reported losses from deepfake scams in 2026 being 263% higher than in 2025.

Other datasets also support the trend of AI being used more frequently by scammers. Chainalysis reported a significant increase in inflows to impersonation scams, while the FBI’s Internet Crime Report recorded a high number of complaints related to AI and cryptocurrency descriptors.

The use of AI by scammers makes impersonation cheaper, more convincing, and easier to scale. Attackers can use AI to maintain conversations in multiple languages, create synthetic videos, clone voices, and generate fraudulent documents and communications.

In the crypto world, where transactions are irreversible once authorized, deepfake scams pose a significant risk. Smart-contract vulnerabilities remain common, but attacks focused on compromising infrastructure and operations have led to the largest losses. Deepfakes help attackers obtain cooperation rather than simply stealing access, making post-onboarding identity checks more critical.

Crypto companies need to implement stronger verification processes to prevent deepfake scams. This could include multiperson approval, pre-established confirmation channels, and delays before new withdrawal addresses become active. Blockchain monitoring tools can help detect suspicious activity, but the key lies in challenging the identity of the person giving instructions before a transaction is signed.

In conclusion, as AI continues to make impersonation more effective, the focus on verifying the authenticity of users before authorizing transactions becomes increasingly important in combating deepfake scams in the crypto space.