What happened
The Five Eyes alliance issued a joint cybersecurity advisory on Tuesday warning that the next wave of frontier AI models will materially lower the cost of offensive cyber operations within months, not years, according to reporting from Crypto Briefing. The advisory, signed jointly by agencies including the US Cybersecurity and Infrastructure Security Agency, the UK's National Cyber Security Centre, the Canadian Centre for Cyber Security, the Australian Signals Directorate, and New Zealand's GCSB, flagged AI-assisted vulnerability discovery, automated phishing at scale, and machine-generated malware as the three threat vectors moving fastest.
It urged operators of critical infrastructure, financial market participants, and digital asset service providers to revisit their threat models now. The agencies did not name specific AI systems or vendors. They did make clear the warning was timed deliberately ahead of model releases expected over the second half of 2026.
Why it matters
Crypto sits in the bullseye. The sector has bled more than $2B to exploits, bridge hacks, and social engineering over the past twelve months, per industry trackers, and the attack surface keeps widening as DeFi total value locked climbs back through pre-2022 highs. AI-assisted offense changes the unit economics of an attack.
A junior operator with access to a capable model can now write convincing English-language phishing aimed at a Coinbase Prime relationship manager, scan a Solidity contract for reentrancy patterns, and draft a credible-looking governance proposal in the same afternoon. None of those tasks previously scaled. They do now.
