Accountability Gap
The distance between the humans whose choices (build, deploy, approve, configure) shaped what an agent could do, and the specific harmful action the agent eventually takes on its own, with no human reviewing that specific action before it happens. The gap widens as autonomy increases and narrows as meaningful human checkpoints are added. The gap can also hide inside a shared responsibility: when several value-chain roles each assume another role covered a safeguard, the result is an action nobody specifically owns, an "everyone and therefore no one" pattern that looks like distributed accountability but functions as none at all.
Defined in 2 GAGE programs, which carry 3 distinct definitions of it. The wording above is taught in Certified Agentic AI Governance Professional (CAAGP).
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 distance between the humans whose choices (build, deploy, approve, configure) shaped what an agent could do, and the specific harmful action the agent eventually takes on its own, with no human reviewing that specific action before it happens. The gap widens as autonomy increases and narrows as meaningful human checkpoints are added. The gap can also hide inside a shared responsibility: when several value-chain roles each assume another role covered a safeguard, the result is an action nobody specifically owns, an "everyone and therefore no one" pattern that looks like distributed accountability but functions as none at all.
the unclear chain of responsibility when an AI agent acts and causes harm; under precedent such as Moffatt v. Air Canada, organizations, not agents, bear legal and professional accountability.
The shift in where responsibility must be engineered once a system moves from producing outputs a human must act on to taking actions on its own, with no human checkpoint at the moment of action. Named here, developed fully in Topic 0.3. (see Topic 0.3)
Where it is taught
The exact lessons this term appears in. The first module of every program is free with a free account.
- Agentic vs. Traditional AI: What Changed · AI Fundamentals, Certified AI Practitioner: Workplace Foundations
- What an Agent Actually Is · Taking the Controls, Certified Agentic AI Governance Professional (CAAGP)
- Why Agentic Breaks Generative Governance · Taking the Controls, Certified Agentic AI Governance Professional (CAAGP)
Where it shows up at work
A definition is where Accountability Gap starts. The lessons above teach it inside the competencies below, which are what job descriptions ask for and what the GAGE credential records. Every link here is drawn from the same careers map that builds a role's readiness check.
Governs AI agents that take actions: tool access, least privilege, interruptibility, cascading actions and accountability for what an agent did.
3 lessons on this page teach it as a primary outcome.
Redesigns roles as AI absorbs tasks, plans reskilling instead of replacement, and supports managers in honest conversations about how work changes.
One lesson on this page teach it as a primary outcome.
Understands prompt injection, data poisoning, model theft, insecure integrations and excessive agent privileges, and the controls that reduce each.
One lesson on this page teach it as a primary outcome.
Roles that hire for it
Each role page shows the full competency set, what the seat decides, and a readiness check against your own evidence.
- Chief Information Security Officer, AI security focusneeds it at depth
- AI Security Architectneeds it at depth · 8 open postings this week
- Chief Technology Officer, AI engineering editionneeds it at depth
- AI Governance Engineerneeds it at depth · 7 open postings this week
- Agentic AI Engineer, orchestration editionneeds working knowledge
- Operations Manager, AI-readyneeds working knowledge
How job descriptions say it
Postings rarely use the textbook word. These are the phrases the careers map reads as asking for the competencies above.
Where the frameworks place it: IMDA Model AI Governance Framework for Agentic AI; NIST AI RMF; OWASP Top 10 for LLM Applications; Google SAIF.
One interview question about it
An agent can send emails and update records. What may it touch, what needs a human, and how do you prove afterwards what it did?
Asked of candidates for agentic ai controls and authorization boundaries. Answering it in your own words is exactly how a GAGE topic is passed, and the credential records the pass where an employer can check it.
If you come from another field
Some backgrounds already carry this competency, and the readiness check gives credit for it.
- Human resources and people operations carries it as work you can show: Role design and workforce plans are your documents.
- Cybersecurity and IT carries it as a claim to back up: Threat modeling and control design transfer; the AI-specific threats are new.
- A computer science or software engineering degree carries it as a claim to back up: Security fundamentals are a course you took, not yet a control you ran.
Terms it appears with
Not an alphabetical neighbourhood: these are the terms taught in the same lessons, ranked by how often they appear together.