Large language models can be used to scam humans, but AI is also susceptible to being scammed – and some models are more gullible than others Large language models can be used to scam humans, but AI is also susceptible to being scammed – and some models are more gullible than others New Scientist – News