Immunefi CEO says AI models lead to surge in crypto security vulnerabilities

Odaily星球日报讯 Mitchell Amador, CEO of bug bounty platform Immunefi, stated at the WAIB Summit that new AI models such as Claude Opus 4.8 and ChatGPT 5.5 are tilting the cybersecurity landscape in favor of attackers, leading to a resurgence in crypto hacking activities in 2026. DefiLlama data shows that in April 2026, illicit actors stole over $634 million from crypto platforms, the highest monthly total since the Bybit hack in February 2025, which resulted in losses of approximately $1.4 billion. 

Amador stated that the crypto industry is in a critical survival period for the next three to four years, until security teams can leverage similar AI models to build code repositories that attackers cannot breach; if the industry adopts more crowdsourced security solutions, this timeline could be shortened to under two years. Anthropic's latest Claude Mythos model, Fable 5, previously raised concerns about accelerating the ability to exploit crypto vulnerabilities;

Anthropic stated that Fable 5 has protective measures in place, directing topics like cybersecurity towards Claude Opus 4.8. On April 19, an attacker siphoned approximately 116,500 re-staked Ethereum rsETH from Kelp DAO's LayerZero-based rsETH bridge, valued at the time at around $290 million to $293 million. Cross-chain protocol LayerZero stated that Kelp DAO's 1/1 decentralized verification network configuration created a single point of failure due to relying on a single verification path to process cross-chain messages. (Cointelegraph)