Responsible AI
The broad practice of ensuring AI is used within a production in a responsible way. Common themes include copyright, data security, transparency, disclosure, and accountability. In the FRAMES context it is the umbrella concept that the AI Ethics Green Paper (ai-ethics-paper) is built around.
Definition (from the paper)
- “The broad practice of putting effort towards ensuring that AI is used within a production in a responsible way. Common themes include Copyright, data security etc.”
- The question is not whether to use AI, but whether one can demonstrate that it is being used responsibly.
Key themes
- ai-copyright - ownership and claimability of AI-assisted output
- ai-transparency - production-level record of AI usage and decisions
- ai-disclosure - explaining AI use to external parties
- ai-accountability - who is liable when something goes wrong
- eu-ai-act - regulatory compliance
- Training and skills (responsible AI training, e.g. ScreenSkills-accredited)
FRAMES practice
- The project explored responsible AI as a constant thread since the proposal (E&R / T&R track: Ethics & Responsibility, Transparency & Risk).
- Tradeoffs were necessary between responsible-AI aspects and final quality (e.g. choosing less-protective tools when the production already owns the rights to its material).