Simple Known Dangers with AI, and Possible Future Dangers Too.

[Jason]

I think the Citrini Research people probably used AI as a source for their article, so it probably is just AI hallucination

Okay, stupid question, where is Citrini Research mentioned? I couldn't find it in the article you quoted, Bill.

[ssfc72]

Jason
2nd paragraph in the Linked article,   "A feedback loop with no brake': how an AI doomsday report shook US markets"

Found an article that reads like sci-fi that is interesting because it caused some stock prices to fall a few days ago. Read most of the original article as well, but it's long. Some of the people who work in the coding business say the time line is too quick. The thesis is that if ai causes high unemployment, the whole economic system we have built is in danger.

Interesting ideas.

https://www.theguardian.com/technology/2026/feb/24/feedback-loop-no-brake-how-ai-doomsday-report-rattled-markets

[Jason]

Jason
2nd paragraph in the Linked article,  "A feedback loop with no brake': how an AI doomsday report shook US markets

Thanks, Bill. Not sure how I missed it. I think maybe I was looking for it in the other article that you linked.

I ran the substack article through chatgpt asking it the probability it was written by AI. It gave a 75-80% chance for the reasons give here:

https://chatgpt.com/s/t_69bece473bb88191920c19c34463236a

It listed a number of reasons of how it came up with the conclusion. I'm listing them here but you'd have to read the link to get more details for each reason. The titles are obtuse.

• Extremely consistent rhetorical structure
• Repetitive conceptual loops (semantic recursion)
• Very broad domain fluency without friction
• Synthetic realism (plausible but unverifiable specifics)
• Length + coherence combination
• High density of polished, "quotable" lines

The irony of using AI to check if something is AI doesn't escape it. But takes one to know one.

I notice as the bottom of the article they also say that the other co-listed article came up with the idea for the post and it was partly written by them. They also list themselves as the author which isn't totally unusually but usually when you co-write an article, if you list one author, you list them both. So they're not claiming they or the co-author wrote an original piece. ChatGPT suggests that it shows the marks of having been written by AI and edited lightly by a human by replacing some common AI-constructed sentences with more humanized ones, and adding some anecodotes.