Question, did anyone check whether the AI was prompted to generate something like this? I don't know anything about TSAM. Was it asked to, "Write a fake piece about the population of Mars"? I agree with most of the concerns you all have raised about AI, but if we didn't even check whether the AI malfunctioned, I'd say all of the comments discussing how "AI will be the death of facts" are just as devoid of facts as the AI, lol (No offense meant to those comment writers. I'm not sure if they missed this or not, but if so, it's an easy mistake to make). Honestly, I'd say most people are wrong from not fact checking their information more often than the current generation of AI is wrong from the kinds of errors AI is prone to on most non-specialized subjects.
Since I didn't see anyone link the original article that Bing pulled the information from, here it is https://tsam.net/what-is-the-population-of-mars/. This editor responsible for this piece on TSAM pretty well only posts these kind of nonsense articles (https://tsam.net/how-cold-is-ice/, https://tsam.net/crocodile-underwater/). It is funny that Bing grabbed one of his nonsense articles, but my guess is that the editor is purposefully prompting the AI to produce nonsense, which doesn't really say much about the factuality of AI.
I looked into the site. From what I could tell and given the timeline, it looked like the person was using an older model of text generation to create this spam at the time. The one about crocodiles being without scales and being slow and ocean dwelling but also being somehow active is the kind of nonsensical zero context preserving stuff gpt used to generate before the gpt 3 model released.
