I attended a digital event, 7 December, where Microsoft launched the Ten Guidelines for Product Leaders to Implement AI Responsibly following their own journey. This is a really useful document and has been collated with diverse perspectives, lived and possessional skills sets. It is where technology meets society and business and research have been working together to enhance the output.
Microsoft shared their path to a responsible AI governance model.
- AETHER - AI & Ethics in Engineering & Research
- ORA - Office of Responsible AI
- RAISE - Responsible AI Strategy in Engineering
The AI guidelines process has 3 stages:
- Assess & prepare
- Design, build, & document
- Validate & support
The report explains the actionable steps
There is a Responsible AI dashboard which is helpful for actionable insights. The responsible AI dashboard includes: Error Analysis Model Statistics, Data Explorer, Aggregate feature importance, What-if counterfactuals, Causal analysis.
There is a Responsible AI Toolbox to get started with