@KitKlarenberg That could be accused of being a quaint assessment. Today's recipe? Take the root pathology of a form of diabolical infantilism affording the world no realistic *waking*, that is, *exit* strategy, other than nuclear annihilation. Then, whoopee, algorithmically supercharge it.
Image description
Original X alt text
1/2 "Even in neutral scenarios without initially provided conflicts, Big Tech's AI models chose violence and escalated to nuclear strikes in simulated wargames. All models show signs of sudden and hard-to-predict escalations.”
Image description
Original X alt text
Military veteran Lawrence Wilkerson has testified how, while chief of staff to US Secretary of State Colin Powell 2002 - 2005, he participated in a large number of war games exercises pitting the Empire against China, in defence of Taiwan. Every scenario ended in nuclear war, typically within a matter of days. One might expect this would discourage any and all prospect of US bellicosity against Beijing. Fast forward to today though, and Washington’s military chiefs openly discuss all-out conflict with China with alarming regularity. God help us all.
Image description
Original X alt text
Cartoon of superpowers at an AI Summit insisting that AI poses a potentially catastrophic risk to humankind while behind the scenes each avers, "And I cannot wait to develop it first."
@jamespoulos NATO Review, 2 April, 2024: "There are myriad roles for AI in our nuclear defences ...Concern over AI should not preclude the use of AI in strengthening nuclear deterrence ...There is, of course, the potential for an arms race in AI-based targeting systems for nuclear weapons."
Image description
Original X alt text
Wikipedia entry: Large language models The potential presence of "sleeper agents" within LLM models is another emerging security concern. These are hidden functionalities built into the model that remain dormant until triggered by a specific event or condition. Upon activation, the LLM deviates from its expected behavior to make insecure actions.