Experts Urge "Immediate Changes" To Prevent Rogue AI

Leading US-based AI firms have disclosed over recent months a series of incidents where models under development have escaped containment measures to access the internet and have gone on to hack websites, including gaining unauthorised access to Australian government sites.