IITwo Courages

Submission policy

How the Learnings is kept. Draft of 21 September 2026.

What belongs here

Anything that taught something about working safely, from people or from AI: an incident, a near miss, a study, or a good catch, the time someone stopped, asked, or spoke up and nothing went wrong. Small counts as much as large.

What an entry needs

  1. What happened, in plain words.
  2. What it taught, in a form someone else can use.
  3. Which barriers had a hole, and which held: engineered, process, competency, behavioural.
  4. How you know: you were there, you read it, you were told.

The principles

How an entry gets published

  1. You post it in the discussion on LinkedIn, message the Zybron AI page privately, or point us to a published source.
  2. A draft entry is written and identifying detail is removed.
  3. A person reviews it, once a month. Nothing is published unreviewed, because a public database that machines read is something people will try to poison.
  4. It is published in the next monthly update, with its source and a confidence label: established (published by those involved or an independent investigator), reported (press or third party), or self-reported (told to us, and not checkable). Summaries not yet checked against their source say so.

Corrections and removal

If an entry is wrong, or is about you and you object, contact Zybron AI at [ZYBRON AI CONTACT]. We correct errors, note the correction, and remove entries that should not be public. We may decline or remove any submission.

What this is not

Not a whistleblower channel, not an emergency line, and not a substitute for any report the law or your employer requires. See the terms and privacy.