Defense in depth
A security and reliability principle, applied here to data governance, in which multiple independent protective layers are stacked so that a single point of failure in any one layer does not by itself allow a violation to reach production; a producer-side contract and consumer-side validation together are an example.
Defined in 4 GAGE programs, which carry 6 distinct definitions of it. The wording above is taught in Certified AI Data Governance Professional (CADGP).
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 security and reliability principle, applied here to data governance, in which multiple independent protective layers are stacked so that a single point of failure in any one layer does not by itself allow a violation to reach production; a producer-side contract and consumer-side validation together are an example.
The security principle, illustrated by the Section 3G password-reset example, of layering multiple, independent controls (structural, then guardrail, then human approval) so that a single failure at one layer does not by itself produce a harmful outcome.
Stacking multiple independent checks (input filtering, the model, rule-based output validation, human review) so no single failure passes straight through to a real-world consequence. The output-validation layer is the highest-leverage and most-skipped.
The engineering practice of stacking multiple, independent safeguards (prevention, detection, mitigation, and recovery layers) between a potential failure and harm, so that any one layer failing still leaves the others standing.
Layering multiple, independent controls around a sensitive asset so that the failure of any single control does not leave the asset fully exposed. The direct response to single point of failure findings like the UnMarker watermark stripping research, which showed one control alone, however sophisticated, can fail.
Where it is taught
The exact lessons this term appears in. The first module of every program is free with a free account.
- The Safety Components · The Agent, Deconstructed, Certified Agentic AI Governance Professional (CAAGP)
- Data contracts: producer and consumer agree in writing, and builds break when they lie · The Catalog That Lives, Certified AI Data Governance Professional (CADGP)
- Governance that survives: rebuilding the estate's defenses so the next attack finds less · Adversarial Data Governance, Certified AI Data Governance Professional (CADGP)
- AI System Reliability and Safety Engineering · Risk, Resilience and Frontier AI, Certified AI Transformation Professional (CATP)
- Security and Adversarial Threats · Risk, Resilience and Frontier AI, Certified AI Transformation Professional (CATP)
- Designing so that one failure does not become a disaster · Failure Analysis, Learning From How Robots Break, Engineering Judgment and Professional Formation
Where it shows up at work
A definition is where Defense in depth 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.
Plans for a vendor failure, an unsafe model change or a suspended service: rollback, manual fallback, data export and safe decommissioning.
2 lessons on this page teach it as a primary outcome.
Designs tests for factuality, robustness, fairness, safety and abuse resistance with rubrics, baselines and thresholds, and says what a score misses.
2 lessons on this page teach it as a primary outcome.
Designs safeguards, family protocols and reporting paths for older adults, minors and people who cannot challenge an automated outcome.
2 lessons 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.
- AI Incident Response Leadneeds it at depth · 1 open posting this week
- Chief Information Security Officer, AI security focusneeds working knowledge
- Chief Risk Officer, AI risk editionneeds working knowledge
- Chief Technology Officer, AI engineering editionneeds it at depth
- Model Risk Managerneeds it at depth
- AI Evaluation Specialistneeds it at depth · 6 open postings this week
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: DORA; NIS2; NIST AI RMF MANAGE 4; NIST AI RMF MEASURE 1 and 2.
One interview question about it
Your AI vendor shuts down next month. What was in your continuity and exit plan, and what did you fail to plan for?
Asked of candidates for ai resilience, continuity and exit planning. 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.
- Cybersecurity and IT carries it as a claim to back up: Continuity and recovery planning are part of the job.
- Model risk, credit risk and quantitative analysis carries it as work you can show: Test design and benchmark selection are the craft.
- A computer science or software engineering degree carries it as a claim to back up: Testing is second nature; AI evaluation design is a new shape of it.
Terms it appears with
Not an alphabetical neighbourhood: these are the terms taught in the same lessons, ranked by how often they appear together.