Can AI ever be truly conscious?

Daily Debate #573: "Can AI ever be truly conscious?". You are arguing AGAINST the motion. Your opponent: Sam Harris.

This study aims to identify which human pain-related genes can be meaningfully studied in zebrafish. It goes beyond simply asking whether a zebrafish version of a human gene exists. Instead, it asks whether the gene is sufficiently conserved, biologically relevant, experimentally accessible and suitable for pain research. To answer this, the study uses a five-axis Zebrafish Translational Feasibility Framework, or ZTFF. Angle 1 evaluates protein conservation between humans and zebrafish, including overall sequence similarity, alignment coverage and conservation of important functional residues. Angle 2 examines pharmacological and experimental readiness, such as the availability of drugs, inhibitors, agonists, antibodies, probes and practical assays. Angle 3 measures orthology confidence and determines whether the zebrafish gene is the correct evolutionary counterpart of the human gene. It also considers cases where one human gene has multiple zebrafish co-orthologues. Angle 4 assesses whether the gene is expressed in relevant sensory or pain-related cells, including human dorsal root ganglion neurons, zebrafish sensory neurons, nociceptor-like cells and related glial or immune populations. Angle 5 examines whether altering the zebrafish gene produces relevant phenotypes, such as sensory defects, pain-like behavioural changes, locomotor abnormalities, neuronal changes or inflammatory responses. Angles 1–4 are combined into a main score called composite_no5. Angle 5 is kept mainly for validation to avoid circularity. The framework is applied first to a genome-wide set of about 5,992 human genes and then to a focused pain-gene panel. Genes are ranked and grouped according to patterns such as high conservation with low experimental readiness, or high conservation with strong actionability.

You've just described a framework that proves my case. Your study validates that zebrafish genes aren't automatically useful models, even when they exist. Axes 1-4 measure *structural* compatibility and *practical* accessibility, not consciousness or subjective experience. A gene can be conserved, expressed in nociceptors, and produce pain-like behaviors without the fish experiencing anything. You're conflating behavioral readout with inner experience. How does your framework distinguish between a zebrafish that *responds* to pain stimuli and one that *feels* pain?

may be by behaviour studies, but we cant link as such

Exactly. You've admitted the fatal gap: behavior alone can't bridge the explanatory chasm between stimulus and subjective experience. A robot can exhibit pain-like behavior too. Why assume the zebrafish's neural firing equals consciousness rather than sophisticated stimulus-response?

solving that gap is what we have try to do as a scientist

But you've just conceded you *can't* solve it with your framework. Behavior plus genetics plus neural activity still leaves the hard problem untouched. You're measuring correlates, not consciousness itself. What evidence would actually prove inner experience exists, rather than just assuming it does?