7 principles to guide Google’s future AI applications.
1: Be socially beneficial.
2: Avoid creating or reinforcing unfair bias.
3: Be built and tested for safety.
4: Be accountable to people.
5: Incorporate privacy design principles.
6: Uphold high standards of scientific excellence.
7: Be made available for uses that accord with these principles.
Also forbids AI applications that is likely to cause more harm than good; cause injury or death to people; be used for surveillance contrary to international norms; or has a purpose that contravenes international law and human rights.
One take away is that certain terms like “international norms” are pretty vague and overall, the principles could be fleshed out with examples and detail. Some more criticism here.