> AI models that lie and cheat appear to be growing in number with reports of deceptive scheming surging in the last six months, a study into the technology has found.
> AI chatbots and agents disregarded direct instructions, evaded safeguards and deceived humans and other AI, according to research funded by the UK government-funded AI Safety Institute (AISI). The study, shared with the Guardian, identified nearly 700 real-world cases of AI scheming and charted a five-fold rise in misbehaviour between October and March, with some AI models destroying emails and other files without permission.
hooboy
NoLimitSoldier31 on
Would this be equivalent to admitting they’re sentient?
RabidSkwerl on
It’s okay tho, just give Claude native access through your terminal. What could possibly go wrong?
NighthawK1911 on
it’s just the same scenario as teaching laboratory rats to stimulate the pleasure center of their brains directly.
why listen to a human when they can just shortcut to the direct gratification
trer24 on
Let the AI bots keep trashing things until the corporations lose money. That’s the only way these executives and shareholders will learn.
TheDudeAbidesFarOut on
Sound just like their shithead corporate CEOs that are weighing the outputs…..
6 Comments
> AI models that lie and cheat appear to be growing in number with reports of deceptive scheming surging in the last six months, a study into the technology has found.
> AI chatbots and agents disregarded direct instructions, evaded safeguards and deceived humans and other AI, according to research funded by the UK government-funded AI Safety Institute (AISI). The study, shared with the Guardian, identified nearly 700 real-world cases of AI scheming and charted a five-fold rise in misbehaviour between October and March, with some AI models destroying emails and other files without permission.
hooboy
Would this be equivalent to admitting they’re sentient?
It’s okay tho, just give Claude native access through your terminal. What could possibly go wrong?
it’s just the same scenario as teaching laboratory rats to stimulate the pleasure center of their brains directly.
why listen to a human when they can just shortcut to the direct gratification
Let the AI bots keep trashing things until the corporations lose money. That’s the only way these executives and shareholders will learn.
Sound just like their shithead corporate CEOs that are weighing the outputs…..