Abstract

The essay aims at discussing the use of artificial intelligence (AI) by parliaments, by distinguishing among four hypothetical uses: to assist in drafting legislation; to create secondary legislation; to crowdsource legislation; to replace parliaments by virtual politicians. It will be argued that while some of these uses do not raise concerns, three promises underlying algorithm democracy cannot be integrated within the framework of liberal democracy. These promises are: that popular will may emerge from data mining; that AI democracy may give rise to a truly successful technocracy untainted by human errors; and that the opposition between democracy and technocracy may be overcome. These promises have been characterized as illusions.