AI Accountability Vocabulary

1234567891011121314
Across
  1. 2. Unfair prejudice in data that leads to discriminatory outcomes
  2. 4. The EU AI Act classifies systems based on this (Unacceptable, High, etc.)
  3. 5. The ability to provide a human-understandable reason for an AI's output
  4. 8. Legal responsibility for damages caused by a product or system
  5. 11. The set of rules or instructions followed by a computer to solve a problem
  6. 13. The obligation to explain, justify, and take responsibility for AI decisions
  7. 14. The moral principles that govern how AI should be designed and deployed
Down
  1. 1. The structural concept of human oversight: In, On, or Out
  2. 2. A system where the internal logic is hidden or too complex to interpret
  3. 3. The landmark European regulation that categorizes AI by risk levels
  4. 6. A formal review to check if an AI system complies with safety standards
  5. 7. The principle of being open about how an AI system is developed and used
  6. 9. A subset of AI based on neural networks that is often difficult to explain
  7. 10. The capacity of an AI to act or make decisions without human intervention
  8. 12. The fuel of AI; if this is biased, the output will be biased too