[ISEA2024] Paper: Yufan Li, Zhan Wang & Theo Papatheodorou — Staying Vigilant in the Age of AI: From Content Generation to Content Authentication

Abstract

Keywords:
Fake Content Generation, ChatGPT, Fake Content Reasoning, Practice-led research, Digital veracity Assessment, AI in Media and Communication

This paper presents the Yangtze Sea project, an initiative in the battle against Generative AI (GAI)-generated fake content. Addressing a pressing issue in the digital age, we investigate public reactions to AI-created fabrications through a structured experiment on a simulated academic conference platform. Our fmdings indicate a profound public challenge in discerning such content, highlighted by GAI’s capacity for realistic fabrications. To counter this, we introduce an innovative approach employing large language models like ChatGPT for truthfulness assessment. We detail a specific workflow for scrutinizing the authenticity of everyday digital content, aimed at boosting public awareness and capability in identifying fake materials. We apply this workflow to an agent bot on Telegram to help users identify the authenticity of text content through conversations. Our project encapsulates a. two-pronged strategy: generating fake content to understand its dynamics and developing assessment tech-niques to mitigate its impact. As part of that effort we propose the creation of speculative fact-checking wearables in the shape of reading glasses and a. clip-on. As a computational media art initiative, this project underscores the delicate interplay between technological progress, ethical considerations, and societal consciousness.

  • Yufan Li, Computational Media and Arts Hong Kong University of Science and Technology (Guangzhou), China
  • Zhan Wang, Computational Media and Arts Hong Kong University of Science and Technology (Guangzhou), China zwang834
  • Theo Papatheodorou, Computational Media and Arts Hong Kong University of Science and Technology (Guangzhou), China