AI cyber: Artificial intelligence systems are advancing so rapidly that they could outsmart existing cybersecurity defences within months, according to a warning issued by the Five Eyes intelligence alliance.
The agencies from the United States, United Kingdom, Australia, Canada and New Zealand said that AI is lowering the barrier for cyber attackers and accelerating the speed and sophistication of digital threats.
In a joint advisory, the group cautioned that long-standing security assumptions are becoming outdated far faster than before, requiring urgent updates to national and corporate cyber strategies.
World Cup 2026 identity: Ayari reflects rise of dual-national stars
The warning follows concerns raised by US AI firm Anthropic, which reported that its advanced models were capable of identifying software vulnerabilities at unprecedented speed.
Security officials said organisations must adopt AI-enabled defence systems, modernise legacy infrastructure and restrict access to critical networks to reduce exposure to attacks.
The advisory stressed that cyber breaches are now inevitable, but rapid detection and response could prevent them from escalating into major operational and financial crises.
Anthropic recently restricted access to its newest models following a US national security directive limiting foreign use, highlighting growing government concern over AI capabilities.
The move also comes amid ongoing debate in Washington over how far AI regulation should go, with some policymakers pushing for tighter controls while others advocate for looser oversight.
Officials warned that the pace of frontier AI development means cyber risks are no longer stable, but continuously evolving in short cycles measured in months rather than years.
