BackDegree of Automation and Control
Category
Risk Domain
AI systems acting in conflict with human goals or values, especially the goals of designers or users, or ethical standards. These misaligned behaviors may be introduced by humans during design and development, such as through reward hacking and goal misgeneralisation, or may result from AI using dangerous capabilities such as manipulation, deception, situational awareness to seek power, self-proliferate, or achieve other goals.
"The degree of automation and control describes the extent to which an AI system functions independently of human supervision and control."(p. 12)
Entity— Who or what caused the harm
Intent— Whether the harm was intentional or accidental
Timing— Whether the risk is pre- or post-deployment
Other risks from Steimers & Schneider (2022) (7)
Fairness
1.1 Unfair discrimination and misrepresentationAI systemUnintentionalPost-deployment
Privacy
2.0 Privacy & SecurityAI systemOtherOther
Complexity of the Intended Task and Usage Environment
7.3 Lack of capability or robustnessAI systemUnintentionalPost-deployment
Degree of Transparency and Explainability
7.4 Lack of transparency or interpretabilityAI systemUnintentionalPost-deployment
Security
2.2 AI system security vulnerabilities and attacksOtherOtherPost-deployment
System Hardware
7.3 Lack of capability or robustnessAI systemUnintentionalOther