Attribution matters, not the AI: economists propose reputation-based proof of human authorship

Afinetheorem · x · 2026-09-30

Economist Afinetheorem endorses John Horton's suggestion to rethink scientific publishing incentives in the AI era. Echoing Ivan Werning, he argues scholars could commit to human-authored work reputationally—with costly punishment for lying—because what matters is reliable attribution of authorship, not the AI itself. Horton adds that science is far older than papers and journals, so academia can design new approaches; Google could build such a verification system "in a hot minute." The discussion points to a pragmatic path: rather than banning AI, make "who wrote what" verifiable.

Related event: Economists Call to Bar AI-Generated Papers from Academic Journals(2 posts)→

Original post →

More from AGI Musings

AGI Musings channel →