AI and democracy – a complicated relationship

I am honest: out of the many podcasts to which I have been invited, this was the toughest one, I would say. Johannes Castner and I discussed what in my opinion is a very dangerous intersection, namely the one of democracy and AI. In a nutshell, I’d say “keep the two things apart from each other”. Here are some of my main points:

  • Democratisation of AI just means “we want to sell”; it’s a misuse of the term democracy in order to legitimize massive roll-out of AI.
  • It’s totally not smart to allow more people to use AI as a powerful decision-maker because powerful decisions are the context where you least should use AI.
  • No, we should not replace judges with AI. If a judge systematically violates rules, she needs to fear to be fired. An algorithm has no conscience and faces no consequences.
  • No, we should not use algorithms for democratic decision-making. Algorithms represent a very binary view of democracy; a yes or no view; but in fact, democracy is at least as much about exchanging arguments (deliberation) as about making decisions. Dissent and compromise are part of any democracy.
  • The courage to use our own understanding is a precondition for democratic societies. If we delegate this to algorithms, it’s the end of democracy. Democracy is not a mouse-click exercise.
  • A lot of the ideas of ‘enhancing’ democracy with AI correspond to the problematic worldview of a radically decentralized libertarian anarchy that cryptocurrency advocates dream about.

These are just some of my key points. You can also watch the whole talk on Youtube or listen to it on Spotify · Apple or here.