What is the goal of a risk assessment in AI?

• A. To adequately assess risks and establish controls to reduce harm to humans or data.
• B. To maximize AI performance regardless of risk.
• C. To document all data usage for compliance only.
• D. To eliminate all risk completely.

Answer: (A)

Risk assessment in AI is about identifying potential harms and evaluating how likely they are and how severe they would be, so you can put controls in place to reduce harm to people and to protect data. The goal is to create a safety and governance posture that minimizes risk while delivering appropriate performance, not to push performance to the expense of safety. Documenting data usage for compliance can be part of governance, but risk assessment focuses on where risk exists and how to mitigate it. Also, you can’t remove all risk; the aim is to reduce risk to an acceptable level and manage residual risk, not achieve zero risk.