Verifiable AI, in plain English
What it means to verify an AI answer, why a council beats one model, how the post-quantum standards work, and what evidence the EU AI Act actually calls for. Honest, technical, no hype.
What “verified AI” actually means — and what it doesn’t
A cryptographic signature on an AI answer proves where it came from and that it wasn’t changed — not that it’s true.
ReadMulti-model AIWhy one AI model is a single point of failure
One model is one training run and one set of blind spots. A council makes disagreement a signal you can see.
ReadPost-quantum cryptographyThe NIST post-quantum cryptography standards, explained
Three finalized standards (FIPS 203/204/205), one forthcoming (Falcon/FIPS 206), and why the transition is a multi-year job.
ReadEU AI ActAI provenance and the EU AI Act: the evidence you actually need
Signed outputs, an append-only log, and a crypto inventory produce evidence that supports your obligations — not a certification.
ReadCryptographyWhat is a Cryptographic Bill of Materials (CBOM)?
A machine-readable inventory of the crypto you actually use — and the honest first step of any post-quantum migration.
ReadVerifiable AIHow to verify an AI answer: a practical guide
If the answer carries a proof, you can check it yourself — in your browser, offline, or from code. Here’s how.
ReadReliabilityAI hallucination: why confident AI is still often wrong
Language models predict plausible text, not verified facts — which is why confidence and correctness are not the same thing.
ReadPost-quantum cryptographyHarvest now, decrypt later: the quantum threat, explained honestly
The urgency isn’t a quantum computer today — it’s that data recorded today can be decrypted the day one arrives.
ReadAI provenanceContent provenance for AI: C2PA, watermarking & signatures
Watermarking, C2PA Content Credentials, and cryptographic signatures each answer “where did this come from?” — with different tradeoffs.
Read