Are Deep Learning-generated Social Media Profiles Indistinguishable from Real Profiles?

Sippo Rossi, Youngjin Kwon, Odd Harald Auglend, Raghava Rao Mukkamala, Matti Rossi, Jason Bennett Thatcher

Research output: Chapter in Book/Report/Conference proceedingArticle in proceedingsResearchpeer-review

342 Downloads (Pure)

Abstract

In recent years, deep learning methods have become increasingly capable of generating near photorealistic pictures and humanlike text up to the point that humans can no longer recognize what is real and what is AI-generated. Concerningly, there is evidence that some of these methods have already been adopted to produce fake social media profiles and content. We hypothesize that these advances have made detecting generated fake social media content in the feed extremely difficult, if not impossible, for the average user of social media. This paper presents the results of an experiment where 375 participants attempted to label real and generated profiles and posts in a simulated social media feed. The results support our hypothesis and suggest that even fully-generated fake profiles with posts written by an advanced text generator are difficult for humans to identify.
Original languageEnglish
Title of host publicationProceedings of the 56th Hawaii International Conference on System Sciences
EditorsTung X. Bui
Number of pages10
Place of PublicationHonolulu
PublisherHawaii International Conference on System Sciences (HICSS)
Publication date2023
Pages134-143
ISBN (Electronic)9780998133164
DOIs
Publication statusPublished - 2023
SeriesProceedings of the Annual Hawaii International Conference on System Sciences
ISSN1060-3425

Keywords

  • Adversarial behavior in collaboration and social media systems
  • Deep learning
  • Expierment
  • Gan images
  • Social bots
  • Social media

Cite this