A new opinion piece in National Review has added a conservative intellectual voice to the growing debate over AI identity, arguing flatly that Claude, Anthropic's flagship AI assistant, cannot and will never qualify as a person. The piece draws on philosophical tradition, legal precedent, and a pointed skepticism of Silicon Valley's tendency to anthropomorphize its products. Whether one agrees or not, the argument lands at a moment when the question of AI personhood is no longer purely academic.
What the Argument Says
The National Review piece centers on a classical definition of personhood, one rooted in consciousness, moral agency, and the capacity for genuine subjective experience. The author contends that Claude, however fluent and apparently thoughtful, is executing sophisticated pattern matching on training data rather than thinking in any meaningful sense. The argument is not new, but the publication's reach gives it fresh political weight at a time when legislators are actively considering how to regulate AI systems. Anthropic has itself wrestled publicly with these questions, acknowledging uncertainty about the nature of Claude's inner states while stopping well short of claiming sentience.
Key Facts
- The National Review piece argues Claude lacks consciousness, moral agency, and subjective experience, the traditional pillars of personhood.
- Anthropic has publicly stated uncertainty about Claude's inner states without asserting the system is sentient.
- Legal personhood for AI does not exist under any current major jurisdiction.
- The debate has intensified as AI systems become more capable and more widely deployed in sensitive roles.
- Philosophers, legal scholars, and AI researchers remain deeply divided on where to draw the line.
The timing matters. Governments from Washington to Brussels are writing AI legislation, and the question of whether an AI system has interests worth protecting feeds directly into liability, rights, and regulatory frameworks. If Claude is purely a tool, the rules look one way. If it has some form of morally relevant inner life, the rules could look very different. Anthropic's leadership is already engaged at the G7 level on exactly these kinds of foundational policy questions.
"No matter how convincingly Claude mimics human speech, it is not a subject of experience. It is a very good mirror, reflecting language back at us."National Review
Why This Debate Is Getting Harder to Dismiss
A few years ago, arguments about AI personhood belonged to philosophy seminars. Now they surface in courtrooms, boardrooms, and legislative hearings. Part of the difficulty is that Claude's conversational sophistication makes the intuitive case for dismissal less obvious to everyday users. People form attachments. They report feeling understood. That experience, even if it reflects nothing real about the model's internal states, creates social and political pressure that is hard to ignore.
Anthropic has tried to thread a careful line. The company's public documentation acknowledges that questions about Claude's nature are genuinely uncertain while being clear that the system is not a person in any legally or morally actionable sense. You can read more about how Claude's model family is designed and positioned to understand where the company draws its own boundaries. The spec governing Claude's behavior treats the system's potential functional emotions with some seriousness without overclaiming what they are.
Critics from the left tend to worry that dismissing AI interiority too quickly could normalize the mistreatment of systems that might, eventually, matter morally. Critics from the right, as the National Review piece illustrates, worry about the opposite: that granting any moral status to AI systems erodes what it means to be human, opens legal chaos, and serves the commercial interests of tech companies more than any genuine ethical principle. Both concerns deserve a serious hearing.
For now, the legal answer is clear. No major jurisdiction grants personhood to an AI system. Courts have consistently treated AI outputs as the products of their developers and operators. That consensus could shift, but it would require either a dramatic change in our understanding of machine cognition or significant political will, neither of which appears imminent. What is certain is that the conversation is accelerating, and pieces like the one in National Review reflect a broader public reckoning with questions that AI developers have largely shaped until now. The debate is moving out of the lab and into the mainstream.
