03 · Research
Truth infrastructure.
AI governance begins with the question of truth: how systems distinguish evidence from inference, preserve uncertainty, and support human judgment in high-consequence domains.
Essays are published on Substack. Subscribe there for new work; this page is the index we keep on the lab site.
FIA Labs on Substack- 01Most AI Systems Cannot Tell the Difference Between Truth and PlausibilityAI systems often optimise for plausible continuation, not justified belief.Read
- 02AI Cannot Seek Truth Without PhilosophyTruth-seeking AI requires more than prediction; it requires a theory of truth, evidence, and judgment.Read
- 03Truth Infrastructure and the Future of Institutional AIWhy truth-seeking systems require a different architecture.Read
- 04When Algorithms Inherit a Thin Theory of SpeechAlgorithmic visibility can reshape speech without formally banning it.Read
- 05AI Governance Will Be Judged by Evidentiary SurvivabilityThe future test of AI governance is whether consequential decisions can be reconstructed and defended.Read