It is becoming increasingly accepted in social scientific thinking that the digital transformation that is currently revolutionising late modern societies has the potential to fundamentally interfere with what is generally understood by the idea of democracy. Accordingly, in the hope of reconciling democracy and digitalisation, calls for democratic regulation of data-based technologies are growing louder. Yet, questioning this conventional way of approaching the nexus between digitalisation and democracy, this paper asks if such regulatory endeavours (in the hope of shielding late modern societies from becoming fundamentally undemocratic) may actually remain no more than an illusion, or mere simulation (Blühdorn, 2013) – as the digital revolution may have in fact rendered modern societies fundamentally 'ungovernable'.
Accordingly, it is argued that whilst in the 1970s, Western democracies were diagnosed as ungovernable because in the eyes of many conservatives, they were perceived as ‘too democratic’, what is rendering them ungovernable today is, by contrast, their observable departure from democratic principles – most of all through the proliferation of the digital technology of artificial intelligence (AI). Accordingly, the paper contemplates how modern conceptualisations of democracy, often supposed the hallmark of modern Western Enlightenment thinking (as it is based upon the collective rationality of free, equal, and, most of all, autonomous individuals), are affected when the autonomous individual is no longer located at heart of decision-making. Where does it leave current understandings of democratic accountability, when the algorithm sits in the driver’s seat, not only figuratively, but soon also in reality? Is it conceivable that the opaqueness of algorithmic decision-making in an emerging "black box" society replaces the market as the central legitimising factor of the late modern political constitution?
Furthermore, technological innovation as a defining feature of the late modern "high-speed society" (Rosa, 2008) operates on profoundly different time frames than lengthy, careful and parliamentary democratic decision-making. As has been the case for instance with social media hate speech, algorithmic regulation generally operates in retrospect only. That this is politically deproblematised, at least by some, is illustrated by the German Free Democratic Party’s (FDP) 2017 federal election slogan Digital First – Bedenken (reflection) Second. Thus, secondly, the relationship between AI and democracy is contemplated with respect to their definingly different operational time scales.
Thirdly, given that we are currently not in fact witnessing further democratisation but instead an autocratic-authoritarian turn (Blühdorn, 2022), this paper finally asks how this development may, in turn, impact current thinking about the future role of AI in late modern societies.
