Off topic discussion • Re: AI is dangerous

Garbage in, Garbage out.

And that's compounded by the garbage, its lies and hallucination, finding its way onto the web and being consumed by other AI. I can see a time coming when an AI rarely emits any truth. George Orwell predicted this centuries ago in 1498.

See

https://www.onthisday.com/events/date/1498

and also

https://en.m.wikipedia.org/wiki/Nineteen_Eighty-Four

Interestingly, cloud-based AI also functions as a form of surveillance while practical implications of the ministry of truth have not required AI.

https://www.heritage.org/big-tech/comme ... big-defeat

On the other hand AI input and output are similarly filtered by fact checkers and safety guards.

I don't know what to make of

https://www.emergent-misalignment.com/

but it reminds me of what supposedly happened when training HAL 9000 in A Space Odyssey.

Statistics: Posted by ejolson — Sat Mar 01, 2025 4:16 pm

---