Consequences of AI for democracy and political participation

This is part of a bigger project at the ISF Munich – Institute for Social Science Research. 

Prof. Dr. Jeanette Hofmann and Dr. Clara Iglesias Keller will focus on the relationship between artificial intelligence and democratic institutions, with special focus on the interaction of these technologies and political participation.

Out of the lessons from the history of technological infrastructures, there are two aspects that hold special importance to this analysis. The first one refers to the ambivalence with which new technologies present themselves, as they arise with potential to emancipation and domination. Alongside promises of economic development and social welfare improvement, their expansion may reconfigure individual and collective autonomy – either by affecting civil liberties directly, or by transforming social practices and institutions that shape individual choices in different ways. The second aspect deals with uncertainty. The changes enabled by new technologies cannot be fully comprehended as of their contemporary deployment. Thus, public discourse and policy choices are likely to contribute to the evolution of AI systems.

Debates on the deployment and regulation of artificial intelligence tend to portray the uncertainty of AI’s impact in terms of risks and opportunities, with the present state of technology and democracy as usual reference point. However, the socially most relevant changes of novel infrastructures are often those that are unexpected and transformative in the sense that they alter the institutions that structure our way of thinking and acting. Such changes are difficult, if not impossible, to predict in terms of risks and opportunities.

In order to understand how the proliferation and use of AI systems may affect political participation, we suggest focusing on indications for transformative changes at the “democratic interface”, understood as the “communication and organization processes that engage citizens with institutions of collective self-governance” (Bennett et al 2018: 1657). In particular, this project will study three perspectives regarding how machine learning techniques relate to political participation: (i) the conditions for political participation; (ii) as means or media to support participation in democratic processes, and (iii) as citizens’ voice in the policy debates currently shaping AI regulation.

The first one – i.e. conditions for political participation – refers to the triad individual autonomy, AI and democracy – where AI alters both individual autonomy (and thus substantive conditions for political participation) and democratic institutions (hence affecting conditions for procedural participation). The latter is also affected indirectly through impacts on autonomy, as limitations on individual rights (e.g. equality and the right to not be discriminated) will also affect access to and exercise of participation mechanisms.

The second concern addresses AI’s promises to enhance individual emancipation. While the public debate focus currently centres on dangers of AI, it should be considered that these systems may also enhance democracy, for example by extending the scope of self-government towards predicted futures.

The third perspective refers to political participation and representativeness as a legitimacy standard for AI targeted policies. Besides stakeholders’ participation in decision-making being a well-known legitimacy standard for regulatory processes (Schmidt 2013), it gains greater importance in a context where policy initiatives are said to shape the very object they intend to regulate. 

The project consists of a literature study on each one of these topics, with a focus on translating them into future policy concerns.