Blessing or curse? A survey on the Impact of Generative AI on Fake News

TL;DR

This survey addresses the rising convergence of Generative AI and Fake News by organizing current work into five clusters: enabling technologies, creation, social-media distribution, detection, and deepfakes. It employs a Structured Literature Review to map literature up to March 2024, highlighting methodological rigor, venues, active groups, and notable gaps. The authors document how GenAI both enables sophisticated fake content and provides detection tools, stressing ethical and societal implications and the need for robust mitigation strategies. The work offers a comprehensive roadmap for researchers, practitioners, and policymakers seeking to preserve information integrity in an era of advanced synthetic media.

Abstract

Fake news significantly influence our society. They impact consumers, voters, and many other societal groups. While Fake News exist for a centuries, Generative AI brings fake news on a new level. It is now possible to automate the creation of masses of high-quality individually targeted Fake News. On the other end, Generative AI can also help detecting Fake News. Both fields are young but developing fast. This survey provides a comprehensive examination of the research and practical use of Generative AI for Fake News detection and creation in 2024. Following the Structured Literature Survey approach, the paper synthesizes current results in the following topic clusters 1) enabling technologies, 2) creation of Fake News, 3) case study social media as most relevant distribution channel, 4) detection of Fake News, and 5) deepfakes as upcoming technology. The article also identifies current challenges and open issues.

Figures (3)

• Figure 1: Structural overview of the domain of Generative AI and its impact on Fake News, illustrating the main themes of creation and detection, along with their subtopics and interrelations. Blue indicates the main themes of the domain; green highlights the subtopics related to these main themes; and yellow denotes key technologies and methodologies central to understanding and addressing the challenges in this domain.
• Figure 2: Schematic representation of the GPT architecture. The model processes input through an embedding layer followed by position encoding to maintain sequence order. This is then passed through multiple transformer blocks, each comprising multi-head self-attention mechanisms and feed-forward neural networks, to generate the output text. The architecture illustrates the flow from input to the generated output.
• Figure 3: Timeline of research activities by leading groups from 2018 to 2024, illustrating the dynamic landscape of contributions in the fields of digital forensics, disinformation detection, and generative AI. Each colored box represents a significant milestone or publication by the respective research group.