DARTS - five foundational pillars
D - Data Governance
Know what information your AI uses and keep it safe and private. Most AI problems start here.
A - Accountability
Be clear on who's in charge of the AI, who intervenes, who answers when it goes wrong. "The AI did it" is not an excuse.
R - Risk Management
R - Risk Management
Anticipate the ways that AI can go wrong; bias, model drift, hallucination, emergent agent behaviour and have a plan for each.
T - Transparency
T - Transparency
R - Risk Management
Clear explanation of how and where you use AI, with testing to prove it's working. More and more, this is expected by the law, too.
S - Stakeholders
T - Transparency
S - Stakeholders
Mechanisms for the people affected by your AI to give feedback and raise concerns. Governance no one can question isn't governance.