Comment AI trained on AI will irreversibly collapse (Score 1) 98
AI trained on AI hallucination will irreversibly and irreparably collapse. That was well-documented here: https://ancillary-proxy.atarimworker.io?url=https%3A%2F%2Fwww.nature.com%2Farticle...
It gets worse when nation-states like russia and China are actively trying to make that happen. We cannot devalue human intelligence and human contact with reality, and we have to whitelist verifiable information. I believe we're going to need to slow down training of the largest models and work on human-legible knowledge bases for highly vetted reasoning agents.
The industry needs a reality check, and ethics need to assert themselves, to find solid ground and lasting usefulness. Throwing vast amount of funding and energy at building foundations on sand is not the way forward.