I remember an SF short story where somebody tried to time travel to the past and ended up way out in space.

[highplainsdem]

21. I'm just going to summarize what happened, because with that long message you just posted from that

chatbot's reply, added to our previous long messages, it would be an impossible long OP in GD. It's easy for DUers to just click on the DU link for this thread, and I'll include the specific numbers of the replies here, including this one you just posted. That way people will have the complete thread for context.

Since that is a chatbot reply, it can't be considered accurate or true, either, unless checked against other sources. Is that Google's Gemini? It offered statistics but didn't provide any sources to check for the accuracy and hallucination rates it gave. In particular, the rate of hallucinations it gives for AI summaries looks very low compared to what I've heard about.

I honestly don't understand what that chatbot meant here:

Demand Context: For high-stakes advice, ask the AI: "Ask me 5 questions about my situation before you give me any advice". If it gives confident answers without needing your specific details, it is likely guessing.

First of all, it's a bad idea to ask a chatbot for high-stakes advice.

And I think what the chatbot meant there was that you should provide it with context. If you were asking for health advice, for instance - though that's particularly risky - and you wanted to know whether a certain medicine might help you, you'd want the chatbot to know about all your health problems and what other meds you take, and supplements. That's true when talking to doctors as well. But that is you supplying information. Doctors - and pharmacists - should also check for possible interactions, and usually do.

And every bit of info you give a chatbot might be exploited by others. No matter what the AI company says, you can't trust the company not to sell your data or use it for training AI. You can't even trust the AI company to keep your private chat sessions off the internet. Chat transcripts from more than one AI company have shown up indexed on Google:

Hundreds of thousands of Grok chats exposed in Google results
https://www.bbc.com/news/articles/cdrkmk00jy0o

ChatGPT Briefly Made Chat Logs Accessible on Google. Yikes.
https://www.vice.com/en/article/chatgpt-briefly-made-chat-logs-accessible-via-google-search-yikes/

Hundreds Of Anthropic Chatbot Transcripts Showed Up In Google Search
People’s conversations with Claude began popping up in Google search results — just like what happened with ChatGPT and Grok. They’ve now disappeared.
https://www.forbes.com/sites/iainmartin/2025/09/08/hundreds-of-anthropic-chatbot-transcripts-showed-up-in-google-search/

Even if your entire chat doesn't show up on Google, part of it might pop up later because it was used as training data.

And you can pretty much guarantee that whatever you tell it will be used later to target you with ads, and to manipulate you through a chatbot or algorithms.

AI companies are all harvesting data. Your data is valuable to them. Chatbots are there primarily to help them, not you.

I'll post here again with the link to the GD message.