Helena Jańczuk

Risks of Using Artificial Intelligence in Creating the Image of Politicians and in Electoral Campaigns

Introduction

Risks of using artificial intelligence in creating the image of politicians and in electoral campaigns. Examine AI risks in political image creation & electoral campaigns. Covers deepfake technology, user profiling, microtargeting, and strategies to minimize these digital threats.

Abstract

In the light of the rapid development of advanced technologies in recent years, many questions have been raised about the future application of available technological solutions in various spheres of life, including politics. An important issue that should be discussed in this field concerns the risks associated with the use of artificial intelligence algorithms in creating the public image of politicians and in electoral campaigns. This paper is based on the concept of eroded epistemics, which is a part of Existential Risk Analysis for AI research. Using the AI Safety Research perspectives of monitoring and systemic safety, it examines the potential risks of using AI in politics and ways to minimize them. The analysis is based on the examples of actions of American politicians. Firstly, the threats of using deepfake technology in creating and manipulating the image of politicians such as Nancy Pelosi, Barack Obama, and Donald Trump, are presented. The second part of the paper discusses user profiling and microtargeting strategies and how they may form opinions and influence voters’ decisions. Finally, examples of present‑day solutions that are being developed to combat these risks are described.

Review

This paper addresses a highly pertinent and critical issue concerning the rapidly evolving landscape of artificial intelligence and its profound implications for democratic processes. By focusing on the risks associated with AI in shaping the public image of politicians and influencing electoral campaigns, the authors tackle a topic of immediate global relevance. The adoption of the "eroded epistemics" concept, derived from Existential Risk Analysis for AI research, provides a solid theoretical foundation, framed within the AI Safety Research perspectives of monitoring and systemic safety. This grounding suggests a rigorous approach to understanding not just the immediate dangers, but also the broader societal impact on truth and knowledge within political discourse. The paper delineates its analysis into several key areas, effectively illustrating the multifaceted nature of AI-driven risks. Firstly, it provides concrete examples of deepfake technology's manipulative potential, referencing well-known American political figures like Nancy Pelosi, Barack Obama, and Donald Trump, which grounds the theoretical discussion in real-world scenarios. Secondly, the discussion of user profiling and microtargeting strategies highlights the subtle yet powerful ways AI can be leveraged to influence voter opinions and decisions, raising important questions about ethical persuasion and informed consent. The inclusion of current solutions being developed to combat these risks is a valuable addition, moving beyond mere problem identification to potential mitigation strategies. While the paper makes a strong contribution to understanding these emerging threats, some areas could benefit from further elaboration. Although the analysis is based on examples of American politicians, a brief discussion on whether these risks are universally applicable or have unique manifestations in other geopolitical contexts would enhance the paper's broader applicability. Furthermore, while "eroded epistemics" is mentioned, the degree to which this concept is thoroughly integrated and analyzed throughout the discussion of deepfakes and microtargeting, beyond its initial introduction, could be strengthened to fully leverage its analytical power. Nevertheless, this is a timely and important contribution to the AI ethics and political science literature, providing valuable insights for policymakers, researchers, and the public alike.

Full Text

Comments

You need to be logged in to post a comment.