SECAI Core · phase 6 of 15
Accountability-AI
An AI principle of who is responsible for decisions and the AI systems that support them at an organizational and individual level
The Explain card
- Plain English
- Accountability in AI is the principle of defining who is responsible for decisions and for the AI systems that support them, at both the organisational and individual level.
- Example
- A hospital deploys an AI triage tool. Accountability means naming who owns the model, who approves changes, and who answers when it mis-prioritises a patient, rather than everyone pointing at "the algorithm".
- Why it matters
- Without named owners, nobody patches the model, monitors for drift or responds to an incident. Clear accountability makes AI systems governable and gives defenders someone to escalate to when something looks wrong.
- Hook
- "The algorithm decided" is not an answer; someone owns the algorithm.
Where it sits in the deck
Phase 6: Model Quality, Ethics, and Responsible AI Principles
Before deploying a model, understand the quality and ethical dimensions — bias, fairness, explainability, and the human oversight needed to govern outputs responsibly.
