AIThe argument against treating AI hallucinations as a solved problem
A critic accuses AI companies of deploying technology as though hallucinations had been fixed, calling the rollout dangerous and negligent.
TLDR
One post calls for renewed attention to AI hallucinations, arguing that the problem was never solved even as companies continued deploying the technology. The author also claims hallucinations cannot be solved with what they call “statistically weighted random number generators,” and characterizes continued deployment as dangerous and negligent.
Combined views
250
1 Source, first seen 2d ago
The argument against treating AI hallucinations as a solved problem
A critic accuses AI companies of deploying technology as though hallucinations had been fixed, calling the rollout dangerous and negligent.
TLDR
One post calls for renewed attention to AI hallucinations, arguing that the problem was never solved even as companies continued deploying the technology. The author also claims hallucinations cannot be solved with what they call “statistically weighted random number generators,” and characterizes continued deployment as dangerous and negligent.