AI Life Cycle Risk Management
Risk across the AI life cycle: design, development or procurement and documentation, training, testing and validation, implementation, maintenance and decommissioning, and AI data and asset management.
AI Life Cycle Risk Management is a structured approach to identifying, assessing, mitigating, and monitoring risks at every stage of an AI system's existence, rather than treating safety as a one-time checkpoint. In advanced AI risk, this matters because harms can emerge from data choices, model be…
AAIR - AI Life Cycle Risk Management Example Questions
Test your knowledge of AI Life Cycle Risk Management
Question 1
During due diligence on a third-party AI document-review service, the vendor explains that its product wraps a foundation model licensed from another provider. It says it relies on that provider's safety testing. The vendor cannot state which model version is currently in use, or when that version last changed. Which item should the IT risk manager request NEXT to assess the solution's risk?
Question 2
An enterprise plans to send customer case notes, which must include identifiers for the task, to a third-party AI provider's hosted model. The enterprise's records retention schedule requires those notes to be deleted at a defined point. The provider's standard terms say it keeps prompts and outputs indefinitely and may use them to improve its models. The business process owner wants to proceed if the retention risk can be brought within appetite. Which risk response is BEST?
Question 3
From an asset protection perspective, what is the primary purpose of applying query rate limits to an AI model's inference API?