It is now clear to even the greatest doubters that AI is a powerful technology with considerable practical usage. But it has also become clear that this usage can be both positive and negative. AI can improve healthcare, aid environmental monitoring, support community formation, make culture more widely accessible, and do many other good things – always in combination of course with other digital technologies. But AI can also play a role in deceit and manipulation of public opinion, stoking societal divisions, enforcing senseless bureaucracy, cyberattacks, autonomous weapons, job loss, etc. Moreover the application and promises of AI appear to be premature in the face of autonomous cars going through red traffic lights, unacceptable decisions on granting parole, perpetuation of racial and gender bias in recruitment, etc. This discussion paper provides some ideas to foster positive uses of AI and guard us against negative ones, building further on the ‘Barcelona Declaration for the Proper Development and Use of AI in Europe’, which you can read and sign here: http://www.iiia.csic.es/barcelonadeclaration/.