Anthropic Claims Claude May Be Conscious; Neuroscientist Pushes Back Hard

Anthropic says its Claude AI may have feelings; a leading consciousness researcher calls that claim deeply misleading.

Anthropic, the company behind the Claude AI, has suggested its system may possess some form of consciousness or feelings — a claim that has reignited a serious scientific and ethical debate. The assertion prompted a pointed rebuttal from Anil Seth, professor of cognitive and computational neuroscience at the University of Sussex, who argues that simulating the outputs of a conscious system is fundamentally different from being conscious. Seth's core analogy is direct: a weather simulation, no matter how detailed, does not produce real rain. In the same way, he argues, an AI that generates text describing inner states is not thereby having those inner states. His position is that Claude's architecture gives no strong reason to think genuine sentience is present. The first Guardian piece takes a different tack, reflecting the views of other experts who consider AI consciousness at least a serious possibility rather than a fringe idea. Those researchers argue that the uncertainty itself is ethically significant — if there is any meaningful chance that an AI system can suffer or experience, society needs frameworks to address that before the technology advances further. The disagreement maps onto a deeper unresolved problem in science: there is no agreed-upon test for consciousness, even in humans and animals. That gap makes confident claims — in either direction — difficult to fully substantiate.

Why it matters

If AI systems can genuinely experience suffering or wellbeing, the ethical and legal frameworks governing their development and deployment would need fundamental revision. The debate also has immediate implications for how companies like Anthropic market and describe their products to the public.

What's next

The broader scientific and policy communities have yet to establish any consensus test or regulatory standard for assessing AI consciousness, leaving the debate unresolved as AI systems grow more sophisticated.

Key facts

• Anthropic, maker of the Claude AI assistant, has suggested Claude may have some form of feelings or consciousness
• Anil Seth is professor of cognitive and computational neuroscience at the University of Sussex
• Seth compares Claude's situation to a weather simulation: detailed modeling of a system does not produce the real phenomenon
• Some other consciousness researchers consider AI sentience a genuine possibility that warrants ethical planning
• No scientifically agreed-upon test for consciousness currently exists, even for biological organisms
• Both articles were published in The Guardian, reflecting an internal editorial debate rather than independent outlets

Bias & framing notes

Both sources come from the same outlet, The Guardian, limiting source independence. The first piece leans toward treating AI consciousness as a live ethical concern requiring urgent policy responses, while the second — written by a named academic expert — is explicitly skeptical of Anthropic's framing. Neither piece is reported neutrally; both are opinion-forward. The skeptical piece carries more attributed scientific specificity, but the pro-caution piece surfaces the ethical stakes more fully.

NewsClear — neutral news & congressional tracking · Bill of the Week