Rewriting History: Is AI Disinformation's Secret Weapon?
Ramsawak, Dominique G.
2023
-
The proliferation of easy-to-use Generative Adversarial Networks (GANs) has enabled the creation of highly realistic and convincing images based on textual descriptions, and raises concerns about the potential misuse of AI-generated images for disinformation and other malicious purposes. This experiment aims to investigate the ability of individuals to differentiate between AI-generated and authentic ... read moreimages, and how this knowledge could be used to mitigate the potential negative social impacts of AI-generated images. Moreover, the study discusses the background of generative AI and its accessibility to society, as well as highlights the prevalence of deepfakes and how the use of AI-generated images can make it more difficult to discern truth from fiction, leading to a situation where people become more skeptical of all images, known as the Liar’s Dividend. Using a commercially available GANs tool, we designed a scenario where individuals were tested on 1) their ability to identify AI generated and authentic images and 2) how confident they were in their decision. We found that individuals are not able to differentiate between AI-generated and authentic images. Lastly, this study offers policy recommendations for governments and other stakeholders, based on the experiment’s findings, that can help mitigate the effects of AI-generated disinformation. It is crucial to ensure there is an awareness of the ability of AI-generated images to be used maliciously, with the aim of preventing the spread of false information, and to maintain the free flow of information for people to make informed decisions.
Submitted in partial fulfillment of the degree Master of Arts at the Fletcher School of Law and Diplomacy.read less - ID:
- pv63gf529
- To Cite:
- TARC Citation Guide EndNote
- Usage:
- Detailed Rights
