This is an overtly most relevant contribution. The debate about ethics in and for AI must keep pace, but it has to be accompanied by institutional processes and actions from the very first moment, which is: now.

It is good to see philosophers leave their ivory towers and have their say.

It is better to see policy-makers pro-actively take up the subject.

It is best to set out for a broad public discourse that is governed not only by interests, but also by wisdom.

Hello! I´m not sure of the convenience of making a global AI charter. In fact, I´m not sure of creating anything global related to AI, even though nowadays we live in an apparently "global village". I think that each country, each culture has their own characteristics, and will assume AI in an specific way depending on that... So, It might be more useful to create specific ones depending on the culture, having into account the others, of course.

While I fully agree with the goals mentioned in the article, however I believe that the recommendations need more discussion.

For the arms, while we have to prevent "relinquish human control almost entirely", Paul Isaac's  already pointed out in his comment below that distinguishing offensive and deffensive weapons has been difficult.

While we should reduce the "risk [...] to dramatically increase the overall amount of in suffering the universe", I think that we should distinguish the biological suffering (resulting from the survival reflex) and a "sufffering" of artificial intelligence used as euphemism for some "human friendly" explanation of the background algorithm. To reduce real suffering, we should probably focus on effects of some AI algorithms on individuals and population (can selection of articles using simple AI keep one in his bubble and lead to suicide? Can preference of fake news based on AI algortithm endanger democracy?) and try to prevent those.

We should work on social cohesion, however I would prefer adaptation of already proved approaches, like anti-monopoly and anti-trust legislation and keep experiments with unconditional basic income as a separate topic independent from AI charter.