While LinkedIn accounts for about one-third of all scanned items, it represents nearly two-thirds of all AI content detected, suggesting a high level of automation in professional networking environments.
The research highlights that longer content is generally more likely to be AI-generated than short-form posts across most platforms.
On X (formerly Twitter), nearly half of all articles were identified as either fully AI-generated or heavily AI-assisted.
In contrast, platforms like Reddit showed lower overall rates of automation, largely because the vast majority of replies remain human-authored, even as top-level posts see increasing AI saturation.
Despite efforts by some platforms to downrank automated "slop," the prevalence of AI writing continues to grow in professional and long-form spaces.
Analysts suggest this trend is driven by integrated platform tools that encourage AI assistance and a growing willingness among users to use automation for professional branding.
The study concludes that undisclosed AI content is becoming a standard fixture of the social media landscape, potentially shifting how users perceive authenticity online.