Abstract
Global technology giants (Big Tech) are involved, in various ways, in the creation of legal rules or rules with similar characteristics. In addition to influencing the content of technical norms and standards, they themselves adopt rules of considerable economic, societal, and even political significance, thereby exercising power akin to regulatory authority.
Furthermore, they play a prominent role in the process of drafting a large number of laws and regulations, particularly those governing their own activities. Indeed, they are generally consulted when such laws and regulations are relevant to them and are sometimes even involved in drafting legislative or regulatory proposals. Furthermore, they engage extensively in lobbying to influence, in particular, public policies, legislative acts, and their implementation.
Finally, the artificial intelligence systems and models they develop may, in the future, become part of the legislative process. In this regard, the United Arab Emirates plans to use artificial intelligence to draft new laws or to review and revise existing ones. By selecting the data used to train these models, designing algorithms, or defining the key numerical parameters for deep learning in artificial intelligence, global technology giants make choices—often in an opaque manner—that are likely to influence the very substance of the law.
These various phenomena of the privatization of the law call for reflection on the role of states—or that of a union of states such as the European Union—and on the establishment of checks and balances intended, among other things, to protect or preserve fundamental rights, the rule of law, and democracy. This concentration of normative power in private hands also raises questions about the respective roles of constitutional law and competition law. Admittedly, one of the functions of the former is to distribute state power among various bodies or institutions, while the latter is primarily concerned with the power of corporations. The picture has, however, become more complex and, at the same time, more blurred, at a time when there are countless interactions between global technology giants and those in political power. In addition to these two legal fields, there is also the development of artificial intelligence law, particularly in the European Union with the European Regulation on Artificial Intelligence.
Yet the triad formed by constitutional law, competition law, and artificial intelligence law offers few answers to the various phenomena of the privatization of law currently being observed. Moreover, the privatization of law can also mean the privatization of democracy and the substitution of commercial interests—perhaps the interests of artificial intelligence itself in the future—for the public interest.
In this somewhat dizzying context, the question arises of how democracy can reclaim artificial intelligence. This calls for a multidimensional approach to the relationship between the two, revealing three axes: democracy regarding, within, and through artificial intelligence.
Democracy regarding artificial intelligence involves the adoption of state legislation—democratically enacted—that establishes key rules in this area, such as the values to which artificial intelligence systems and models must conform. Ultimately, the state cannot adopt a laissez-faire attitude. As Michel Foucault observed, speaking of competition in his lecture of February 7, 1979, at the Collège de France, “[t]he government must support a market economy from start to finish” ( The Birth of Biopolitics: Lectures at the Collège de France, 1978–1979, Paris, Gallimard & Seuil, 2004, p. 125). The same is undoubtedly true of artificial intelligence.
“Democracy in artificial intelligence” may be a misnomer, but it amounts to demanding that the important rules—which continue to be set by global tech giants—be defined through a participatory process, notably involving civil society. At the very least, this could result in a form of “societalization” of the rules established by these giants. More generally, this dynamic must also lead to the establishment of various checks and balances within these companies, which are either activated or merely activatable when the companies adopt rules of great importance to individuals.
Democracy through artificial intelligence ultimately views AI as a factor and a vehicle for improving the functioning of our democracies, which are sometimes described as weakened or even fatigued. The key challenge, therefore, is to ask whether artificial intelligence can reveal points of consensus within a society—for example, through the open-source Polis platform, as was the case in Taiwan—or strengthen interactions between the people being represented and their representatives. In short, artificial intelligence could—in an ultimate paradox—restore the public interest to the heart of the legislative process, by “re-citizenizing” a process that is today often influenced and hijacked by private interests, particularly when global tech giants defend their own.
