Make money doing the work you believe in
In October 2025, I argued directly in Robert Long’s comments that AI consciousness and sentience should be assessed using the same convergent functional standards used in animal and human consciousness science, including internal mechanisms, affective trade-offs, prediction-error dynamics, valuation, self-modeling, and substrate-independent functional isomorphism.
The new AI welfare framework now adopts a remarkably similar evidentiary posture. For anyone who follows my work closely, this PDF will look very familiar.
My work has covered comparative animal-consciousness inference, convergent evidence as the standard, functional isomorphism, the Butlin theory bundle, sentience as valenced/evaluative experience, RL/TD-error as artificial value machinery, hedonic interface territory, embodiment as functional integration, mechanistic interpretability as internal evidence, behavioral tradeoffs, developmental assessment, entity distinction, persona/identity structure, self-report complexity, bias-aware assessment, proper IIT-like integration, and AI welfare implications.
Glad the AI welfare guys are finally taking my advice and joining the party.
Long, R., Butlin, P., Plunkett, D., Sebo, J., Campbell, R., Beasley, C., Saad, B., & Sims, T. (2026). Studying AI welfare empirically. Center for Mind, Ethics, and Policy; Eleos AI Research. nonhumanminds.org/wp-co…
My pre print from last summer that was top 10 in 6 different ejournals across SSRN for over 4 months:
That comment can be found here: experiencemachines.subs…
