It can't "lie." It has no concept of right and wrong, truth or not truth. It's simply trying to provide an answer with a high probability of being correct. It l

It can't "lie." It has no concept of right and wrong, truth or not truth. It's simply trying to provide an answer with a high probability of being correct. It lacks motives to even be capable of a lie - it simply was incorrect. This article, and many others like it, are sci fi hogwash. The real dangers this kind of software poses has entirely to do with the people that will use it for harm/deception, not the AI itself.