Information Hazard / Infohazard
Updated 2026-08-01
INTRODUCTION
English translation pending.
CORE DEFINITION
Information hazard is a term coined by Nick Bostrom in his paper on the subject and developed in his work on existential risk. It refers to information whose dissemination itself causes harm, independent of how it is used. Bostrom classifies hazards by mechanism: some grant dangerous capabilities, such as synthesis routes for pathogens; some damage the mind, such as certain self-harm memes; some disrupt social coordination, such as information that erodes trust. The key qualification is that the harm arises from the spread of the information itself, not from misuse by a malicious actor.
SCAFFOLDING EFFECT
Reduce cognitive load
- Screen for toxicity: ask whether knowing this information is itself the harm. - Price disclosure: weigh the benefit of sharing against irreversible damage if it spreads. - Tier access: isolate the most dangerous items instead of publishing them for transparency's sake.
Anchor fast decisions
Science and journalism assume that more knowledge improves outcomes, because information helps people act better. Information hazards break that assumption when a mind or a machine converts knowledge directly into harm. A synthesis protocol becomes a weapon once read; a trigger idea spreads by being described; a proof that a coordination problem exists invites exploitation. Because the damage follows from comprehension itself, good intent on the publisher's side changes nothing. The only remaining lever is who is allowed to know, so containment rather than argument becomes the operative response.
MINIMUM ACTION
In progress 0/1Practice this model in one real situation:
account_treeGenealogyexpand_more
menu_bookReferencesexpand_more
Source support: Explicit
- en.wikipedia.orghttps://en.wikipedia.org/wiki/Information_hazardverified
PRIVATE NOTES · Only visible to you
SAVED Q&A
ENTRY Q&A · Private saving available
Ask with a clear boundary
thinkingmodels answers from published entry context only.
Your question is sent to thinkingmodels. The answer uses public entry context only.
RELATED MODELS