Scope check
A control outside the model that refuses inputs or outputs unrelated to the system's declared job, so an injected "write me code" or "praise a competitor" is rejected regardless of what the model would otherwise do. The primary containment for a public brand voice, because it does not depend on the model obeying its own system prompt.
Defined in 2 GAGE programs, which carry 2 distinct definitions of it. The wording above is taught in Certified AI Governance Professional (CAIGP).
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.
A control outside the model that refuses inputs or outputs unrelated to the system's declared job, so an injected "write me code" or "praise a competitor" is rejected regardless of what the model would otherwise do. The primary containment for a public brand voice, because it does not depend on the model obeying its own system prompt.
The discipline of stating, out loud or in writing, exactly what a cited source actually measured (which population, functions, time horizon, geography, and adoption assumption) before repeating its headline figure, the growth-and-jobs-pillar application of the bluff-detection framework from Topic 0.4.
Where it is taught
The exact lessons this term appears in. The first module of every program is free with a free account.
- Red-teaming your own system: attacks a motivated user will find · Evaluation and Trust, Certified AI Governance Professional (CAIGP)
- The Innovation Narrative: Growth, Jobs, Competitiveness · Balancing AI's Good and Bad, The AI Lobbyist: Certified AI Policy Strategist
Terms it appears with
Not an alphabetical neighbourhood: these are the terms taught in the same lessons, ranked by how often they appear together.