Can people prove they are human without revealing who they are?

Anonymity is valuable online but has always enabled deception, and increasingly capable AI sharpens the dilemma: two trends — AI's growing indistinguishability from people (lifelike content, avatars, agentic activity) and its scalability (cheap, accessible) — threaten to let AI-powered deception overwhelm the internet. Personhood credentials (PHCs) are the proposed countermeasure: privacy-preserving digital credentials, issued by trusted institutions (governmental or otherwise), that let a user demonstrate they are a real person — not an AI — to online services without disclosing any personal information. PHCs need not be biometric and can be local or global. They counter three harms: sockpuppets (fake "people"), bot attacks (automated abuse evading suspension), and misleading agents (AI misrepresenting whose goals it serves).

The keeper is the design tension it resolves: how to preserve anonymity and trustworthiness simultaneously — by proving personhood without proving identity. It reframes the deepfake/bot problem from "detect AI content" to "credential humans."

This is a society/identity anchor with a strong post angle. It is the infrastructural response to the deception landscape documented in How much of the internet is AI-generated now? and Why do AI posts get likes without inviting conversation?, and it intersects the disclosure/identity dynamics of Do humans learn to prefer AI partners over time?.