Who Rules in the Age of Global AI? Liberalist Premise of AI Systems and its Challenge to the Rule by the People

Information systems utilising Artificial Intelligence (AI) have become ubiquitous on a global scale, while also bringing with them societal challenges that call for governance and regulation. Meanwhile, the providers of AI systems decide on how their systems work. They decide both the values reflected by their models, as well as who are represented in their datasets. They choose the weights in decision-making algorithms, and judge what content is to be censured. As these systems are increasingly used in contexts that shape societies and people's lives, the calls for subjecting decisions made by AI providers to public scrutiny have intensified. The markets deciding upon the functioning of AI systems are global, and hence the systemic separation of democratic governance and the liberal market challenges the idea of rule by the people – the demos. As AI systems are currently treated as products rather than systems subject to democratic decision-making despite their arguably political nature, we seem to approach a system of a vote per dollar, where knowledge presents itself to market participants who then vote for the best outcomes with their wallet. This arises from the fundamental premise of liberalism that is guaranteed through supranational law and trade treaties, which makes the question about governance of AI systems largely borderless. In this context, democratic control that has traditionally been based on national conception of demos faces a significant challenge. In this paper, we inspect the liberalist philosophy behind the current AI development that exceeds national borders and how that challenges the idea of democratic AI. We look at the compatibility between AI and democracy through the lens of demos – the people that should have the power over collective decisions that shape people's lives and societies.