Journalists are reluctant to admit they use AI. Not because it's unethical — but because there's no way to prove how they used it.
In practice, reporters quietly use AI all the time: transcribing long interviews with speech-to-text, searching for information, researching background, planning, taking notes. Responsible journalists then verify the facts themselves, make every editorial decision themselves — the result is 100% human creative work.
But if someone asks "did AI write this?" — what's the answer? "Yes, but only for transcription and research" sounds like an excuse. There's no record of where AI stopped and the journalist started.
The fix
RE::DACT records that boundary. Every AI interaction — what was requested, what was generated, what the journalist changed, rejected, or accepted — is hashed with SHA-256 (the same standard used for digital signatures in banking) and recorded on the Hedera hashgraph. The result is a human oversight certificate: tamper-proof, independently timestamped, impossible to falsify after the fact — even by us.
Why it matters legally
A regular log on a newsroom's server is a file that can be edited — in court, it's not evidence. A hash on a public ledger is different: anyone can independently verify that the record hasn't been altered, the same way a notary certifies a document.
When the first accusation of "AI-generated journalism" turns into a lawsuit, a newsroom with a RE::DACT certificate responds with proof, not promises.
The EU AI Act will require this kind of transparency for high-risk AI applications by 2026. The newsrooms building audit infrastructure now are the ones that won't panic later.