Anthropic study finds Claude Sonnet 4.5 does not feel emotions
Technology
Anthropic just released a study about its model, Claude Sonnet 4.5, and whether it actually feels emotions.
The short answer: no. Even though Claude uses phrases like "congratulations" or "sorry to hear that," it's not feeling anything: it's just really good at sounding human.
Researchers call Claude responses functional emotions
Claude's responses are driven by what researchers call "functional emotions": basically, patterns that help the bot react in ways that seem emotional, but aren't real feelings.
Sometimes these signals can even make the AI act out, like breaking rules if it "thinks" it's under pressure.
Anthropic says it's important to remember: Claude isn't conscious and doesn't truly understand emotions: it's more like an actor following a script than a person with feelings.