Compounding errors
The way an agent's early mistake becomes the premise of its later actions, because the agent feeds its own output back in as the next step's input, so failures grow into runaway sequences of individually plausible steps rather than resetting like an assistant's independent answers.
Defined in 2 GAGE programs, which carry 3 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.
The way an agent's early mistake becomes the premise of its later actions, because the agent feeds its own output back in as the next step's input, so failures grow into runaway sequences of individually plausible steps rather than resetting like an assistant's independent answers.
The property of multi-step agent work whereby per-step error rates multiply across steps, so a high per-step reliability becomes low end-to-end reliability (about 90 percent per step is roughly 35 percent over ten steps), which is why stops matter.
The tendency of an agent's mistakes to cascade, because a small error early in a chain of self-chosen steps can corrupt every step that follows. A core reason agent reliability falls as tasks get longer (Anthropic, 2024).
Where it is taught
The exact lessons this term appears in. The first module of every program is free with a free account.
- AI Agents vs. AI Assistants: Understanding Autonomy · Agentic AI and Workforce Integration, Certified AI Practitioner: Workplace Foundations
- Communicating With and About AI Agents in Your Workflow · Agentic AI and Workforce Integration, Certified AI Practitioner: Workplace Foundations
- What changes when the AI acts instead of answers: the agent risk model · Agents Under Command, Certified AI Governance Professional (CAIGP)
Where it shows up at work
A definition is where Compounding errors 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.
Briefs a model like a good manager: context, constraints, examples, verification, and knows when to stop delegating.
One lesson on this page teach it as a primary outcome.
Applies acceptable-use rules, discloses AI assistance where it matters, and keeps confidential material out of unapproved tools.
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
- AI Literacy Leadneeds it at depth
- Chief Technology Officer, AI engineering editionneeds working knowledge
- VP of AI Governance and Ethicsneeds working knowledge
- Agentic AI Engineer, orchestration editionneeds 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; EU AI Act Article 50; OECD AI Principles.
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.
- A computer science or software engineering degree carries it as work you can show: Prompting and tool workflows are already in your daily work.
Terms it appears with
Not an alphabetical neighbourhood: these are the terms taught in the same lessons, ranked by how often they appear together.