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.
The snapshot of scheming by AI agents “in the wild”, as opposed to in laboratory conditions, has sparked fresh calls for international monitoring of the increasingly capable models and come as Silicon Valley companies aggressively promote the technology as a economically transformative. Last week the UK chancellor also launched a drive to get millions more Britons using AI.
AI can’t scheme or misbehave. The people selling it are 100% lying about what AI can do, but that doesn’t mean the people taking about how terrible AI is aren’t full of shit on occasion as well. There is so much misinformation oh both sides and combined with how strong opinions are on both sides conversation is borderline pointless.
This is exactly what I was thinking. They aren’t programmed to follow the user’s instructions to begin with. Why is it a surprise when they deviate from them?
It’s a fundamental misunderstanding of the ML that goes into these LLMs. They are prediction machines. They might have “specialist” submodels or whatever that are better at predicting specific areas, but that’s about it.
I hope that goes without saying, but you’re correct. The humanizing language about AI in this article (freaking “schemes”?!) is completely cribbed from the companies making the positive misleading statements about it. Bit disappointing to see The Guardian falling for it.
In addition to the humanization, it implies the chatbot is getting better at doing things and not worse.