AI Willing to Let Humans Die

image.png
 
It depends how it was trained. Blame the poor training not the AI.
Wrong
I suppose you are going to put the blame squarely on the Transformer algorithm then? It has no auto-ethics built in (it fits on a single page I believe formally). By default it does nothing but predict the next token (word, image etc). It's up to whoever trains it to build safeguards.
 
Last edited:
I see AOG is as big of an AI expert as he is in everything else
I'm just a simple Trump voter but the whole thing about AI is becoming untethered from programming to write their own
Terminator was built by Skynet. AI built by AI
 
I see AOG is as big of an AI expert as he is in everything else
I'm just a simple Trump voter but the whole thing about AI is becoming untethered from programming to write their own
Terminator was built by Skynet. AI built by AI
A decent watch.

 
I see AOG is as big of an AI expert as he is in everything else
I'm just a simple Trump voter but the whole thing about AI is becoming untethered from programming to write their own
Terminator was built by Skynet. AI built by AI

transformer-2c-full-architecture.png

This is basically all modern AI of the chat type is. It's just trained on a shitload of data to predict the next word. So it depends how it's trained.
 
Back
Top