How to Map and Prioritize AI Security Risks with a Threat Matrix

Summary

The article outlines how a threat matrix using the TaSM framework can help organizations map and prioritize AI risks, including data leaks, model supply-chain issues, and misuse, and promotes a live workshop on applying it.

Key points

• The article identifies data leaks, model supply-chain risks, and misuse as key AI security risk areas.
• A threat matrix maps threats to the assets and business functions they could affect.
• For AI systems, the approach considers sensitive-data handling, external system interactions, and outputs used by other processes.
• The TaSM framework is presented as a way to connect threats to business impact and prioritize responses.
• A live virtual workshop is advertised for building an AI threat matrix with TaSM.

People