[Digital Today reporter Yoonseo Lee] A warning has emerged that hacking using artificial intelligence (AI) could significantly worsen security in the cryptocurrency industry over the next year.
Cryptopolitan, a blockchain media outlet, reported on Tuesday that Diogo Monica, co-founder of Anchorage Digital, said at the Solana Breakpoint conference in London that the next 12 months would be a harsh period for crypto security.
Monica said that in the past 3 to 6 months, AI tools have reached a level where they can carry out cyberattacks without human control. Even someone with no knowledge only needs to type, "Please hack this website," targeting a specific site, he said.
The warning comes amid recent major hacking incidents. Michael Coates, chief information security officer (CISO) at the Solana Foundation, said cryptocurrency stolen over the past decade exceeds $13 billion, and more than $7.5 billion of that was concentrated in 2026. The biggest loss from a single incident was the Binance attack, at about $6.3 billion.
Coates warned that as AI sharply lowers the barrier to hacking, small cryptocurrency projects that had been outside hackers' focus could also become targets.
In June, Mitchell Amador, chief executive officer (CEO) of security firm Immunefi, pointed out that advanced AI models have brought about a catastrophe. He said such models worked to attackers' advantage and are leading to more DeFi project hacks in 2026.
Coates and Monica advised crypto companies to start by checking their basic security systems. They particularly stressed protecting employee devices and private keys. If private keys are leaked, attackers can move cryptocurrency elsewhere.
Major exchanges are seeking AI-based defensive measures. Coinbase and Binance were reported to have tried in April to secure access to 'Claude Mythos,' an AI model Anthropic provides on a limited basis. Anthropic explained that the model can independently find and exploit previously unknown vulnerabilities in operating systems and web browsers.
From major exchanges to small and midsize projects, a key task has emerged: preventing a breach of a single employee device or developer account from escalating into fund outflows. A core focus of security checks is who can access private keys and how far privileges are exposed when one account is compromised.