Crypto security firm Trugard and onchain trust protocol Webacy have developed an innovative AI-powered system to combat the increasingly sophisticated tactic of crypto wallet address poisoning. This new tool, announced on May 21st, leverages a supervised machine learning model trained on live transaction data and on-chain analytics. The system boasts a 97% success rate in detecting known attack scenarios, according to Webacy’s co-founder Maika Isogawa, who highlighted that address poisoning is a significant yet underreported scam in the crypto space. This deceptive scam involves attackers sending tiny amounts of cryptocurrency from an almost identical address to a target’s actual address. The aim is to trick users into copying and using the attacker’s address in future transactions, leading to financial losses. In a study conducted earlier this year, over 270 million poisoning attempts were recorded across BNB Chain and Ethereum between July 2022 and June 2024. With success rates as high as $83 million, the attack remains a significant threat. Trugard’s Chief Technology Officer, Jeremiah O’Connor, explained that this new system utilizes the firm’s cybersecurity expertise from traditional systems to adapt for Web3 data. This is achieved by applying algorithmic feature engineering techniques that are commonly used in Web2 and Web3 applications. Most existing Web3 attack detection systems rely on static rules or basic transaction filtering, but this AI-powered system excels in adapting to evolving attacker tactics. Webacy’s team employs machine learning for a dynamic, adaptive solution. The AI system generates synthetic training data to simulate various attack patterns before being trained through supervised learning where it learns the relationship between input variables and the correct output. O’Connor explained that the model is continuously updated with new data as new strategies emerge, while Isogawa emphasized that AI can detect patterns beyond human analytical capabilities.