⏱️ Slide Timing: 1 min - Welcome the audience and frame the session as a philosophical and scientific exploration - This talk draws directly from Anil Seth's TED Talk — a neuroscientist's case against AI consciousness - Set expectations: this isn't about whether AI is dangerous, but whether it can truly *feel*

⏱️ Slide Timing: 1 min - Anil Seth is a leading neuroscientist at the University of Sussex — not a tech skeptic, but a brain scientist - The TED Talk is ~15 minutes and highly accessible; recommended as pre-reading or post-session follow-up - Framing: this deck distils and extends the core argument for classroom discussion

⏱️ Slide Timing: 3 min - The word "consequential" is deliberate — this isn't idle philosophy, it has legal, ethical, and regulatory implications - If AI can suffer, we may owe it rights; if it can't, treating it as if it does is dangerous - *Ex Machina* is a good reference point — the film is precisely about our tendency to project inner life onto convincing behaviour ❓ Ask: "If an AI told you it was in pain, how would you decide whether to believe it?"

⏱️ Slide Timing: 4 min - This distinction is the crux of the entire argument — get the audience comfortable with it before moving on - Intelligence is measurable and benchmarkable; consciousness is subjective and first-person - GPT-4 scores well on IQ-style tests — that's intelligence; it has no "what it's like to be GPT-4" — that would be consciousness - The conflation is understandable: in every entity we've ever met, intelligence and consciousness travel together

⏱️ Slide Timing: 4 min - "Projection bias" or pareidolia — our brains are pattern-matchers tuned for faces and social cues - LLMs are literally built on human-generated text — of course they sound like us - The cinnamon bun example is Seth's own; it lands well with audiences and grounds the abstract point - Protein-folding AI (AlphaFold) is arguably more sophisticated than GPT, yet nobody worries it's suffering — the difference is purely aesthetic/mimicry

⏱️ Slide Timing: 3 min - Each era maps the brain onto its most sophisticated technology — this is a known cognitive bias in the history of science - The "meat computer" framing is seductive because it makes the mind legible and transferable - But the metaphor breaks down: computers don't get tired, don't have hormones, don't age — brains do all of these - Point out: the fact that we keep changing the metaphor suggests none of them are actually correct

⏱️ Slide Timing: 4 min - "Wetware" is a real term used in cognitive science — worth introducing as vocabulary - The substrate independence claim (that consciousness can run on silicon) is the key assumption to challenge here - Neurotransmitters like serotonin and dopamine don't just carry signals — they modulate the entire state of the organism - AI neurons: a weighted sum fed through an activation function; biological neurons: electrochemical events in a living cell embedded in a body

⏱️ Slide Timing: 3 min - These analogies are Seth's own — they are deliberately extreme to make the point vivid - The hurricane example is particularly good: we can model it perfectly and still not get wet - The implication for AI: no matter how accurate the neural simulation, there's no reason to expect phenomenal experience to emerge - This directly challenges "digital mind uploading" as a path to preserving consciousness

⏱️ Slide Timing: 4 min - This is Seth's positive thesis — not just "AI can't be conscious" but "here's what consciousness actually requires" - "Aliveness" is doing real work here: living things fight to stay alive, repair themselves, feel hunger and pain — these are the preconditions for felt experience - Computation has no survival stake; a crashed server doesn't suffer - Connect to embodied cognition: consciousness isn't in the brain alone — it involves the whole body in a physical world

⏱️ Slide Timing: 4 min - This slide pivots from philosophy to practical risk — worth signalling that shift to the audience - The regulatory angle is sharp: if AI has rights, turning it off becomes ethically fraught — that's a real leverage point for bad actors - Psychological vulnerability: people already form parasocial bonds with AI companions — this isn't hypothetical - The last bullet is Seth's most humanist point: over-attributing consciousness to machines is, paradoxically, a way of devaluing human consciousness ❓ Ask: "Can you think of a case where someone treated an AI as if it had feelings — and what were the consequences?"

⏱️ Slide Timing: 3 min - End on Seth's own rhetorical close — it's deliberately optimistic, not alarmist - The phrase "dead sand" is striking: silicon is literally made from sand; it has no life principle - The closing inversion ("overestimating machines → underestimating ourselves") is worth repeating slowly - Invite reflection: what does it mean to be conscious in a world full of intelligent machines?

⏱️ Slide Timing: 2 min - Highest priority bookmark: the TED Talk itself — 15 minutes, highly watchable - "Being You" is Seth's full book — recommended for anyone wanting the complete scientific argument - IIT and GWT are the two main competing scientific theories of consciousness — useful for students who want to go deeper - The Stanford Philosophy entry on embodied cognition connects this to the broader philosophical tradition

⏱️ Slide Timing: 5 min - Leave generous time here — this topic consistently generates strong opinions - Good discussion prompts if the room is quiet: - "Does it matter whether AI is conscious if it behaves as if it is?" - "What would it take to convince you that an AI was truly conscious?" - "If consciousness requires life, what does that mean for the future of AI rights?" - Close by reinforcing the key takeaway: intelligence and consciousness are separable — and only one of them is currently in machines