Robert Long, Executive Director of Eleos AI, a nonprofit that researches AI welfare, argued that intelligence and consciousness in AI should be treated as distinct, and that the simple equation of "smarter equals more conscious" does not hold.
AI Consciousness & Welfare · Eleos AI
"Smarter" Is Not "More Conscious"
Eleos AI director Robert Long warns that intelligence and consciousness may correlate — but they are not the same. Scaling a model up, he argues, should not be assumed to switch consciousness on.
3–4
people at the Eleos AI nonprofit (as of 2026)
19
co-authors on the 2023 paper on consciousness in AI
4
expert stances that sharply diverge on the question
Where the experts stand — from doubt to belief
Robert Long
Cautious, evidence-based
David Chalmers
Keeps it open
Geoffrey Hinton
May already be conscious
Column height reflects how open each figure is to AI consciousness — a spectrum, not a consensus.
From theory to empirical test
Neuroscientific theory
→
Welfare policy
→
Empirical evaluation
Praised
An empirical, practical approach — running dialogue experiments on loneliness, memory continuity and a wish to end conversations — that grounds the debate in evidence and resists excessive anthropomorphism.
Cautioned
There is still little basis to judge whether today's LLMs are genuine welfare subjects — something Long himself acknowledges is only weakly established.
The debate has intensified since 2023 — and it's now moving inside AI companies.
Firms including Anthropic have launched consciousness and welfare research programs, and the second Eleos Conference on AI Consciousness and Welfare is set for Berkeley, September 18–20, 2026.
Continue reading The rest of this article is for AI News Blitz readers. Choose an option below to keep reading.
Already purchased? Sign in ✓ Signed in — this article isn’t included in your current plan.Unlocking the full article…