Enrolment
This course is currently free to enrol in.
Public course
Enterprise AI — Course 17: AI Security and Threat Analysis
Trace AI attack paths, judge enforced controls and prioritise repairs using supplied evidence and explicit authority.
AI securitythreat analysisprompt injectionpermissionsoutput handlingincident response
By Telari Labs
Enterprise AICourse 17 of 25
Learning on Telari
Created by Telari Labs, delivered through Telari
The course creator controls the material and assessment standards. Telari provides structured lessons, practice, feedback against those standards and a record of your progress. See how Telari works.
AI access
The first managed lesson is included on Free. Continue with Plus or your own supported API key; premium assessments and certificates may require Plus.
Audio
No reviewed lesson audio is currently advertised for this course.
Lesson list
- Assets and trust boundariesIdentify what needs protection and trace a supported path from an actor’s influence to a consequence.
- Prompt injection and instruction authorityDistinguish task content from attempted instructions and separate behavioural evidence from enforced boundaries.
- Disclosure and unsafe interpretationFind avoidable data exposure and assess how a receiving component handles generated output.
- Tools and confused authorityResolve differences between a user’s rights, a model proposal and the identity executing an action.
- Data, model and dependency exposureEvaluate persistence and changed dependencies before carrying old security evidence forward.
- Controls linked to attack pathsMatch controls to steps and design tests that observe the intended enforcement.
- Residual exposure and responseSeparate containment from recovery and define evidence-based restoration within stated authority.
- Threat review and prioritiesTurn supported paths and control evidence into a practical repair order and bounded use decision.
- Final assessment — Security decisionsApply the course’s reasoning to six independent supplied cases.