Let’s be clear: LLMs don’t feel pain. And they don’t deserve moral standing. This new paper reports a series of experiments identifying pain…

Let’s be clear: LLMs don’t feel pain. And they don’t deserve moral standing. This new paper reports a series of experiments identifying pain-related representations in LLMs. The authors also show that manipulating these representations changes behavior: some fine-tuned models select a “relief” button less often after it removes the manipulation than when it leaves it in place. These are interesting results and the authors acknowledge that their experiments do not establish conscious experience and could instead reflect role-play. But they also say that “pain-like states would inform debates on AI moral standing and welfare.” And the paper has been widely shared among advocates of AI rights and moral standing. So, let’s be very clear: These findings do not show that LLMs experience pain. Similarities in internal representations and behavior do not imply similarities in subjective experience. A system can represent pain without that representation being painful. A system can simulate rain without getting wet. That’s it. Don’t conflate surface similarity with deep similarity. * Paper in the first reply

Related

• breaking: author of the study that so many “AI influencers” are taking as evidence that LLMs feel pain tells me “we don’t claim they feel pa…
• excellent thread on that “pain” paper
• Let me say this clearly: LLMs cannot feel emotions. Emotions are evolutionary mechanisms. They push us to avoid danger or approach what is b…

Source: Gary Marcus (X) | 2026-09-21