There have been a handful of posts that are clearly either purely copypasta straight from AI, or inspired by AI hallucinations.
Anyone familiar with the Project Skyscraper timeline thus far can spot them a mile away, since they often contain information or concepts that are completely alien or out of left field to the current knowledgebase.
But people who are new may be easily led down the wrong paths, and since there are so many channels (multiple discords, reddit, Atlas-65), this information gets cross-shared and it can easily make its way into the collective investigation effort (Atlas forbid a YouTube influencer takes off with something).
I’m posting this as a sort of caution against relying exclusively or even heavily on AI for your research, and especially to do our best to refrain from posting pure AI generated ideas.
AI can be helpful in fleshing out your own investigative paths, or to help word things better, but as far as relying on it for raw deep investigation–“ChatGPT, tell me what you find in this XYZ web page”–it’s best not to.
It doesn’t understand what it’s saying, it just gives you the statistically most likely best next word, one after another, in the context of your input (or prompt).
This is a great idea for a thread, thanks for posting this. I would give you hearts but I have hit my current limit. I will return in force with them. Here’s some IOU’s in the mean time <3 <3 <3
This ARG is new, chatbots cannot know anything about it. They only know about other, similar ARGs and games. If you give it an unknown image or text, and ask it to “decrypt” something, it will tell you a nice fairytale. Chatbots do not admit to not knowing something…
On the other hand, you can ask it for explanations, like:
Somebody mentioned “rot13” today, what does that word mean?
And then use what you learned. This way, an ARG is way more fun, and you feel like a detective–and not like the manager of an insane asylum.
Exactly, AI can be super helpful for quickly learning things you don’t understand, or getting feedback on an idea you had. Even then, it helps to verify anything you do learn with a secondary source (google still exists)!
And here’s our AI summary brought to us by ETARC AI:
YourBasicMaths initiates a discussion warning the community against relying heavily on AI for solving Project Skyscraper. They note that AI-generated content often contains hallucinations or concepts alien to the ARG’s current knowledge base, which can mislead new investigators and contaminate collective efforts across platforms like Discord and Reddit. While AI can assist in refining phrasing or fleshing out ideas, it should not be used for raw deep investigation because it merely predicts statistically likely words without true understanding…
