In the YouTube video “Why AI Will Never Be Conscious,” Alex O’Connor interviews Anil Seth and asks a number of questions surrounding consciousness and artificial intelligence. I didn’t find the conversation to be all that interesting, but maybe it’s because I’m not the target audience. Regardless, what I did find rather interesting was a comment left by @kevinharrison2074:
“…a system which forms a temporally extended, self-modelling, world-inhabiting process whose internal organization gives rise to an actual point of view”
This is actually good because it provides a rough definition of what subjectivity involves and where it comes from. Sure, “internal organization” is vague but it’s a start. As such, it gets added to The List.
What I will give Seth credit for is his brief mention of autopoiesis at 0:25:06 and his half-hearted attempt at moving away from a computational perspective of consciousness.1 Characterizing consciousness, even loosely, as a controlled hallucination2 is not a helpful why of explaining consciousness though, given the semantic baggage attached to ‘hallucination’. This phrasing puts the cart before the horse, because hallucinations are erroneous perceptions. Perceptions themselves aren’t necessarily erroneous, they are interpretations generated from sensory data.
Maslow’s Law of the Instrument can be applied to research and theory: if the only thing you study is the brain, everything looks like a neuron.
Works Cited
1 Alex O’Connor, Why AI Will Never Be Conscious – Anil Seth, 2026, 1:42:27, https://www.youtube.com/watch?v=lsi8T_WtLnE.
2 Alex O’Connor, Why AI Will Never Be Conscious – Anil Seth, at 1:25:13.