With the popularity of generative AI, social media is being flooded with AI-generated content. According to a recent study by data analysis company Pangram, one out of every four long-form social media posts exceeding 250 words is generated by artificial intelligence. Among these, the professional networking platform LinkedIn has become the "absolute king" of long-form AI spam content.

QQ20260713-143513.jpg

The data comes from Pangram's Chrome extension, which scanned over 1 million posts across five major social media platforms between April and June 2026. The data shows that although LinkedIn accounts for only one-third of all scanned samples, it contributes nearly two-thirds of the detected AI content. On the LinkedIn platform, as many as 41% of long-form posts are marked as AI-written, and this proportion reaches 30% in short-form content (50-250 words), both being the highest among all platforms.

Following closely behind is Medium, with AI proportions of 28% for short-form content and 31% for long-form content. X (formerly Twitter) also shows significant AI involvement in long-form content, with nearly half (29% fully generated, and even higher if AI-assisted content is included) of long articles being generated or assisted by artificial intelligence, but the AI proportion in short-form content is only 9%.

In contrast, some platforms have maintained a high level of original human content. Substack has the lowest proportion of AI-generated long-form content, at around 10% (12% for short-form). Reddit, on the other hand, exhibits unique ecological characteristics: while its individual posts contain a relatively high frequency of AI text (13% for long-form and 3% for short-form), as many as 98% of user replies on the platform are still manually written by real people.

It is reported that the Pangram3 detection model used in this study has a higher accuracy in identifying human-written content, with an error rate of only 0.01%, meaning the actual proportion of AI-generated content on social media may be even higher than the statistical data suggests. Currently, facing the rampant AI-generated content, LinkedIn itself seems to have felt the pressure and has started taking strict measures to combat such posts.