Adversarial manipulation
A failure mode in which a bad output is deliberately induced by an attacker, through a jailbreak that talks the model past its own guardrails or a prompt injection hidden in content the model reads, rather than arising on its own. Distinct from ordinary fabrication because the trigger is chosen and repeatable by anyone who finds it. Deep treatment and defenses belong to Topic 4.3.
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 failure mode in which a bad output is deliberately induced by an attacker, through a jailbreak that talks the model past its own guardrails or a prompt injection hidden in content the model reads, rather than arising on its own. Distinct from ordinary fabrication because the trigger is chosen and repeatable by anyone who finds it. Deep treatment and defenses belong to Topic 4.3.
Deliberately causing an AI system to misbehave by crafting its inputs, most commonly through prompt injection (input designed to override the system's instructions). A rising threat as AI systems are connected to real tools and data.
Where it is taught
The exact lessons this term appears in. The first module of every program is free with a free account.
- The license to govern: explaining to a skeptic exactly how your model fails · Build Before You Govern, Certified AI Governance Professional (CAIGP)
- AI Risk Management · Governance and Responsible AI, Certified AI Transformation Professional (CATP)
Where it shows up at work
A definition is where Adversarial manipulation 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.
Keeps the living record: each risk with a named owner, rating, treatment, residual risk, monitoring metric, threshold and review date.
One lesson on this page teach it as a primary outcome.
Turns controls into enforceable obligations: data use, change notice, audit rights, incident duties, subcontractors, IP, exit and deletion.
One lesson on this page teach it as a primary outcome.
Tiers vendors by use and impact, requests evidence instead of promises, tests in the customer's context, and plans monitoring and exit.
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.
- Third-Party AI Risk Analystneeds it at depth
- Chief Risk Officer, AI risk editionneeds it at depth
- AI Risk Managerneeds it at depth · 2 open postings this week
- Cybersecurity Risk Analystneeds working knowledge
- AI Vendor Risk Managerneeds it at depth
- Third-Party Cyber Risk Managerneeds it at depth
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: NIST AI RMF MANAGE; ISO/IEC 42001 Clause 6; EU AI Act Article 9; GDPR Article 28.
One interview question about it
How do you keep an AI risk register from becoming a list nobody reads? What makes a risk entry actionable?
Asked of candidates for ai risk register and treatment tracking. 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.
- Internal audit and IT audit carries it as a claim to back up: Remediation tracking is familiar from follow-up work.
- Legal and paralegal work carries it as work you can show: Contract terms are your product.
- A law degree (JD, LLB or LLM), no AI work yet carries it as a claim to back up: Contract terms that move risk are a first-year subject.
Terms it appears with
Not an alphabetical neighbourhood: these are the terms taught in the same lessons, ranked by how often they appear together.