Guardrails & Accountability

The less a human watches, the more the guardrails must carry the load. This module is the safety and governance counterweight to the autonomy taught everywhere else in the course.

Theme: Operations & safety →

Topics

5 topics — scope and references for each.

Layered defense — model / app / tool / human
Guardrails work in layers, each catching what the others miss: model-level safety for content, application-level validation for domain errors, tool-level per…
Human-in-the-loop — pre / post / conditional approvals
The patterns for inserting human judgment: pre-action approvals before execution, post-action reviews for quality assurance, and conditional approvals trigge…
Least-privilege & approved-API whitelisting
Defining what an agent physically can and cannot do: whitelisting approved APIs and enforcing least-privilege access so that even a hijacked agent has a smal…
Audit trails
Comprehensive logging as the foundation of accountability: every agent action recorded with who triggered it, what the agent did, what data it used, and whic…
Compliance frameworks — ISO 42001, NIST AI RMF, EU AI Act, GDPR/HIPAA
The regulatory landscape agents increasingly operate in, and how technical controls map to it: risk assessments, audit logging, data minimisation, and human-…

Topics are combined ad hoc for corporate programs and workshops — there is no fixed public duration. Tell us your goals and we design the track.

Discuss a custom program