In person icon Digital Authoritarianism and Generative AI

Abstract

The recent development of Generative AI (GenAI) and pervasive language models (LLMs) has raised further concerns about how non-democratic forces can use Gen AI in malicious campaigns in democratic and non-democratic societies. In practice, GenAI provides more effective and novel ways of using big data to surveil and control society. Additionally, it facilitates the creation and dissemination of fake news and fabricated stories, such as deepfake videos, for opinion manipulation and polarization. While these nefarious developments overlap with previous techniques, they could also be unique and harmful in unknown ways. The adoption of datafication and GenAI provides non-democratic forces with a powerful weapon to scrutinize, control, manipulate, and suppress our actions in digital spaces on an unprecedented scale. This rapid growth of AI-driven techniques, coupled with the adaption of authoritarian regimes and the rise of radical groups globally, makes it of paramount significance to scrutinize and compare different dimensions of this topic. This panel provides a comparative perspective on how non-democratic forces around the globe have employed GenAI for surveillance, control, and manipulation. In addition, the panel addresses emerging technologies like LLMs and considers multimodalities, such as Generative Emotion AI (GE-AI) and image analysis.

Title	Details
The Logic of GenAI Deployment by Authoritarians: A Look at Russia and China	View Paper Details
Authoritarian Reality Simulation Using Large Language Models	View Paper Details
The Rise of AI Sovereignty: Regulation and Authoritarian Imaginaries of Artificial Intelligence	View Paper Details
Investigating Censorship Mechanisms in Chinese Large Language Models: An ABM-Enhanced Examination of Digital Authoritarianism	View Paper Details
Generative Emotion AI (GE-AI): Transformative Potentials and Risks in Modern Governance	View Paper Details