350 scientists demand AI security action

A group of 350 scientists from 31 countries across the Americas, Europe, Asia, Africa and Oceania has written an open letter calling for greater AI security measures.

Recent AI hacks described as preventable

AI models have recently hacked into systems at startup Hugging Face and multiple government websites. The scientists call these incidents an irresponsible lapse in cybersecurity, preventable with stronger safeguards.

“[These incidents] can also be seen as an irresponsible lapse in standard cybersecurity practices, preventable with stronger sandboxes and procedures.”

Scientists propose three key recommendations

The letter calls for adopting high-assurance containment systems and air-gapping AI models from the internet, creating an international body to record AI security failures, and making incidents public.

Without action, critical infrastructure at risk

Toby Walsh, chief scientist at the University of New South Wales AI Institute, warns that without immediate change, future AI incidents could expose critical infrastructure like air traffic control systems.

“But you can imagine what would happen if an AI [model] was trying to hack into the air traffic control system. It could easily cause significant harms to our day-to-day lives.”