Verify a member Contact Member sign in
International Initiative for Augmented Security

Hands-on labs

Every IIAS program builds practical skill through graded labs. There are 27 across our five programs, each set in a realistic scenario.

How labs work

Labs put you in the role of a practitioner at a fictional organization, with real tasks to complete and decisions to justify.

Realistic scenarios

Each program follows one fictional organization, with its systems, data, people and pressures.

Simulated environments

Labs run in your browser on infrastructure the initiative controls. There is nothing to install.

Consistent AI tools

AI tools in labs are simulations that behave the same way for every learner, so results are fair and comparable.

Graded on reasoning

Objectives are checked on completion, and written work is scored against a published rubric.

Lab catalog

All 27 graded labs, by program.

AI Security Fundamentals

Foundation | 4 graded labs
  1. Map the attack surface of an AI assistant

    Identify the data sources, tools, users and integrations of a sample assistant, and where each could be abused.

  2. Harden an AI assistant for daily work

    Review an assistant's settings, judge everyday uses against safe-use rules, and recognize and respond to signs that it was influenced by content it read.

  3. Spot synthetic media and verify a high-risk request

    Assess audio and video samples for synthetic indicators and apply a verification procedure to a payment request.

  4. Classify AI use cases by risk

    Assign risk levels and required controls to a set of AI use cases for a fictional organization.

Synthetic Media and Fraud Defense

Foundation | 5 graded labs
  1. Assess audio, video and image samples

    Examine a set of genuine and synthetic samples, record indicators, and explain where detection is uncertain.

  2. Verify a high-risk request

    Handle an urgent request from an apparent executive, choose trusted contact details, and apply a verification procedure before acting.

  3. Redesign a payment-approval workflow

    Identify weaknesses in a fictional organization's payment process and redesign it to resist impersonation.

  4. Review onboarding cases

    Review remote loan applications, identify likely synthetic identities, and recommend fair, well-reasoned decisions.

  5. Plan an impersonation awareness exercise

    Design a realistic, ethical awareness exercise with clear objectives, reporting routes and follow-up.

LLM and Agentic Application Security

Intermediate | 6 graded labs
  1. Harden an assistant's trust boundaries

    Remove secrets and unsafe rules from a system prompt, separate untrusted content from instructions, and contain what the assistant can do if it is manipulated.

  2. Secure an assistant's output handling

    Map where model output lands, review the code that handles it, and apply encoding, query templates, schema validation and safe rendering.

  3. Harden a RAG pipeline

    Set access and ingestion rules for each knowledge source, enforce permissions at retrieval, review the ingestion queue and respond to poisoned content.

  4. Scope and harden an over-privileged agent

    Redesign an agent's tools, identity, approval steps and limits so it can do its job and nothing more, then test the design against proposed actions.

  5. Design and test a layered guardrail

    Place input, policy and output controls in layers, measure configurations against test results, and choose one that meets the requirements.

  6. Threat model and secure a release candidate

    Map design review findings to the OWASP Top 10 for LLM Applications, prioritize them, fix the release blockers and make the release decision.

AI Governance and Risk Management

Intermediate | 6 graded labs
  1. Draft an AI policy and accountability matrix

    Write an AI policy for the organization and assign roles and decision rights for its AI systems.

  2. Build an AI system inventory

    Identify and document the organization's AI systems, including third-party and embedded AI, and classify each by risk.

  3. Assess risks and populate a risk register

    Assess the risks of selected AI systems and record ratings, owners and treatments in a risk register.

  4. Conduct an AI impact assessment

    Assess the effects of a high-risk AI system on affected individuals and document mitigation decisions.

  5. Design an AI management system control set

    Select and justify controls aligned to ISO/IEC 42001 and integrate them with existing security controls.

  6. Prepare and review audit evidence

    Assemble evidence for an internal audit, identify nonconformities and recommend corrective actions.

AI-Augmented Security Operations

Intermediate | 6 graded labs
  1. Build and tune an ML-assisted detection

    Train an anomaly detection model on authentication logs, tune its threshold and measure false positives.

  2. Triage an alert queue with LLM support

    Use an LLM to summarize and prioritize alerts, then verify each conclusion against the underlying evidence.

  3. Investigate an incident with an AI assistant

    Work through a multi-stage incident using AI-generated hypotheses and queries, and document the findings.

  4. Design an automated playbook with approval points

    Classify response actions by risk and build a playbook that requires analyst approval where needed.

  5. Build an AI performance dashboard

    Calculate precision, drift and override rates from historical data and present them for SOC leadership.

  6. Detect manipulation of a SOC AI assistant

    Identify injected instructions hidden in log and ticket data, and apply controls that contain them.

Start building practical skill

Choose a program aligned to your role. Completing every graded lab is required for your program certificate and the membership pathway.