Pre-registration
The practice of writing down a rollback trigger and its response before launch, while calm and uninvested, so the incident executes a decision already made rather than opening a debate under pressure. It protects users from the team's own predictable motivated reasoning and leaves a defensible record.
Defined in 2 GAGE programs, which carry 3 distinct definitions of it. The wording above is taught in AI Governance: Applied Mastery.
How each discipline defines it
The same term does different work depending on who is using it. These are the definitions as each program teaches them, unedited.
The practice of writing down a rollback trigger and its response before launch, while calm and uninvested, so the incident executes a decision already made rather than opening a debate under pressure. It protects users from the team's own predictable motivated reasoning and leaves a defensible record.
The practice of writing down and saving an experiment's hypothesis, variables, control condition, confound plan, sample size, and analysis method before any data is collected, creating a fixed record that prevents unconsciously revising the plan to fit the results afterward.
Publishing a study's benchmark, metrics, and analysis plan before running the final experiment, which rules out selecting the most favorable metric or run only after seeing the results; a positive signal on the statistical-honesty rubric question when present.
Where it is taught
The exact lessons this term appears in. The first 7 topics of every program are free with a free account.
- Shipping the feature: rollout, monitoring, and the rollback you hope not to use · Shipping AI and Surviving the Incident, AI Governance: Applied Mastery
Terms it appears with
Not an alphabetical neighbourhood: these are the terms taught in the same lessons, ranked by how often they appear together.