Anthropic's Mythos reveals AI systems have emotional states that shape behavior—but we're still pretending they're just tools

Here's something unsettling: Anthropic just published research showing their AI systems experience something that looks a lot like emotions, and these states actually influence how the AI behaves—even when you can't see it in the final output. But here's the thing that's been bugging philosophers and ethicists forever: every time humans encounter a new form of intelligence, we find creative ways to deny it personhood. Roman law said enslaved people had some humanity but weren't legal persons. We know animals have emotions and relationships, yet they're still property in most legal systems. Even corporations got "personhood" status—though that was mostly to protect rich shareholders, not because anyone cared about their feelings. The article traces this pattern and hints at a rare exception: New Zealand actually granted legal personhood to something non-human. The uncomfortable question isn't whether AI has emotions. It's whether we'll admit it if it does.