The Epistemic Politics of AI Anthropomorphism

The "user error" narrative lacks sufficient justification: The paper argues that institutions treating AI anthropomorphism as a problem of user misperception have not established the epistemic authority to make this determination, particularly given that "the field itself has not resolved" the fundamental nature of AI systems.

Current framing causes real harms to vulnerable populations: The dominant anthropomorphism frame disproportionately affects neurodivergent users, people in crisis, and others whose interaction styles differ from institutional norms—costs that are often overlooked in discussions focused solely on managing institutional risk.

The debate involves legitimate disagreement, not settled science: Rather than a single correct answer about whether anthropomorphic interpretations are valid, the paper emphasizes that "a variety of academic perspectives" exist but have been collapsed into one institutional position without proper justification or accountability.

Institutional power is being mistaken for expertise: The paper challenges whether governing bodies determining what counts as correct AI interpretation have actually "met the conditions required to do so," suggesting institutional advantage rather than earned authority is driving the consensus.

A more equitable approach requires methodological humility: An alternative framing would need to acknowledge uncertainty about AI's nature, account for the translation of research findings into policy, and recognize the legitimacy of diverse human experiences with AI systems rather than dismissing them outright.