(02-17-2023, 01:53 PM)The Real Marty Wrote:

(02-17-2023, 12:01 PM)NewJagsCity Wrote: As someone who has worked in IT since 1987, I can say without reservation that this is stupid.  Really stupid.  Anthropomorphism at its dumbest.  I hope the chat box suicides by powering itself down.

Do you think it's possible to create a human being out of a machine?  By which I mean, in your opinion, is it possible to create a machine with enough human characteristics that it is not distinguishable from an actual human?

All of these AI chat bots are completely controlled and parameters or curated data sets are used to get these results. It's simple to give them different personalities, different ages, intelligence, style, viewpoints, etc. They can make that random or they can make it react to your inputs. The main thing is they control what data it sees and how it views that data.

They can influence or even write responses for the human stuff and once you start asking about that, it gives you their answers or it can rephrase it. It's all done by training the AI to respond how you want. That's required less now with the data sets but you still have to tell it what the truth is.

Right now they can fake it for a while and you probably wouldn't notice much. It just depends on how long you talk. If you go into it knowing you aren't talking to a human, you will see the issues. If you are in a situation where you are just chatting or texting with someone, it may be hard to figure out in short interactions.


Until they actually open these up to the internet and let them ingest it's own data and judge the truth, these aren't really AI. That would be the scary side as you now could have differences in every AI depending the configuration and the data it has ingested. It also could be controlled by just putting data online.

Sent from my SM-T970 using Tapatalk

---