3/
Confusing Outputs for Awareness:
Opponents view citing an LLM's self-reported probability of consciousness (e.g., asking a model if it is conscious and analyzing its reply) as a fundamental misunderstanding of next-token prediction, treating sophisticated pattern matching as genuine internal experience.
Gemini says:
Jeff Jarvis’s reaction points to a major point of friction in contemporary AI discourse:
Distraction from Immediate Harms:
Critics argue that debating whether a statistical language model has "feelings" or deserves rights diverts attention away from tangible, present-day issues like algorithmic bias, copyright infringement, energy consumption, and labor exploitation.
