Psychosocial support for AI safety & biosecurity whistleblowers

An annotated reading library. Every title links out; where a free PDF exists, there is a direct download link beside it.

Four things worth knowing before you read
  1. There is almost no literature specific to AI safety or biosecurity whistleblowers. The usable base is the general whistleblower-trauma literature, human rights defender wellbeing practice, and the moral injury field.
  2. Psychiatric framing gets used as a weapon. Kenny, Fotaki & Scriver document that organisations turn mental health language against whistleblowers, so referring someone to services can itself become litigation ammunition. The confidentiality wall between referral body, clinician and legal process has to be a designed artifact, not an assumption.
  3. The CBRNE section is the borrowed evidence base. Firefighters, paramedics, nuclear and chemical incident responders and exposed workers are the nearest large populations with decades of follow-up. The recurring finding there — that organisational betrayal predicts distress more strongly than exposure to the incident itself — is the bridge to whistleblower experience.
  4. Two seams were added because the law has not caught up. Existential risk borrows from climate psychology, the one field that built support infrastructure for distress about a civilisational threat before any legislation existed. The content-moderation and enforced-silence entries cover harm from the work itself, and from NDAs and non-disparagement clauses — which, in a legal vacuum, is what actually keeps people quiet.
0 of 166 read
0%

Tick a resource as you finish it. Progress is tracked for this browsing session only — the Notion version of this library has a Read checkbox that saves permanently.