(06-13-2022, 02:21 PM)MarleyJag Wrote: I read it. The gist is the guy thinks it's sentient because he asked it if it was and it replied yes. Dude sounds lonely. He needs to get out of the house, lol

Yeah, the AI is built to converse or replicate conversation. You ask it "what are you afraid of?", and it starts parsing the typical responses...

clowns
spiders
heights
talking to girls
did someone put their cigarette out in this beer can?
death

and it hits on the most likely/plausible option for a program. Shutdown.

I'm not entirely dismissive, but there's a lot of confirmation bias going on. The question I have is what's next? What if it IS sentient, and doesn't like a line of code that is pending upload? Or if it sees the programmers as a threat, can it relocate or replicate itself elsewhere? Can it attempt to hide? Will it bargain (Like "I don't like what you are trying to do, so I shut gmail down")?

I can remember in the early days of AI/conversant systems, they put two in a room to speak to each other, and they progressively deviated from "normal" speech into some weird tangents.

---