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

TL;DR

This study investigates whether advanced deep learning-generated social media profiles and content are indistinguishable from real ones, finding that even humans struggle to detect fake profiles created with state-of-the-art AI.

Contribution

The paper provides empirical evidence that AI-generated social media profiles are highly convincing and difficult for humans to distinguish from real profiles, highlighting detection challenges.

Findings

Humans often cannot reliably identify AI-generated profiles.

Deep learning can produce social media content that closely mimics real profiles.

Detection of fake profiles remains a significant challenge.

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.