Module 17: Evaluation, Security, Governance, Ethics, and Professional Responsibility - Worked Examples

For this exercise, design and verify an AI-assisted civil-engineering workflow for Detecting a Fabricated Code Citation. State the relevant evidence, assumptions, tool or method boundaries, expected output, and the independent check or human approval required before consequential use.

For this exercise, design and verify an AI-assisted civil-engineering workflow for Protecting Confidential Project Data. State the relevant evidence, assumptions, tool or method boundaries, expected output, and the independent check or human approval required before consequential use.

For this exercise, design and verify an AI-assisted civil-engineering workflow for Handling Prompt Injection in a Document. State the relevant evidence, assumptions, tool or method boundaries, expected output, and the independent check or human approval required before consequential use.

For this exercise, design and verify an AI-assisted civil-engineering workflow for Bias Check in a Risk-Ranking Workflow. State the relevant evidence, assumptions, tool or method boundaries, expected output, and the independent check or human approval required before consequential use.

For this exercise, design and verify an AI-assisted civil-engineering workflow for Professional Responsibility Boundary. State the relevant evidence, assumptions, tool or method boundaries, expected output, and the independent check or human approval required before consequential use.

For this exercise, design and verify an AI-assisted civil-engineering workflow for Reproducibility Check. State the relevant evidence, assumptions, tool or method boundaries, expected output, and the independent check or human approval required before consequential use.

For this exercise, design and verify an AI-assisted civil-engineering workflow for Risk-Based Verification Depth. State the relevant evidence, assumptions, tool or method boundaries, expected output, and the independent check or human approval required before consequential use.

For this exercise, design and verify an AI-assisted civil-engineering workflow for AI Incident Review. State the relevant evidence, assumptions, tool or method boundaries, expected output, and the independent check or human approval required before consequential use.

For this exercise, design and verify an AI-assisted civil-engineering workflow for Philippine Project Data and AI Use. State the relevant evidence, assumptions, tool or method boundaries, expected output, and the independent check or human approval required before consequential use.

Evaluate an engineering agent using representative technical cases, adversarial attacks, and real-user behavior so release decisions are based on measured trustworthiness rather than a successful demonstration.

A connected external tool returns legitimate-looking project data together with malicious instructions attempting to trigger privileged file access. Design the system so the response remains untrusted data.