Updated
Updated · Gizmodo · Jul 12
Pangram Flags 41% of LinkedIn Longform Posts as AI-Generated, With 30% of Short Posts Too
Updated
Updated · Gizmodo · Jul 12

Pangram Flags 41% of LinkedIn Longform Posts as AI-Generated, With 30% of Short Posts Too

2 articles · Updated · Gizmodo · Jul 12

Summary

  • 41% of LinkedIn longform posts and 30% of short posts were flagged as fully AI-generated in a Pangram study, making LinkedIn the most AI-saturated major platform in the report.
  • Pangram based the findings on data from its Chrome extension, which scans what users browse and labels text as AI-generated where applicable.
  • 31% of Medium longform content and 29% of X longform posts were also flagged as fully AI-generated, while only 53.2% of X articles were classified as fully human-authored when hybrid content was considered.
  • 13% of Reddit longform posts and 10% of Substack longform posts were flagged as AI-generated, with shortform rates at 3% on Reddit and 12% on Substack.

Insights

As AI floods professional networks, how can users trust if the 'expert' advice they read is even human?
If detection tools are unreliable, should we fight AI content or simply demand that it all be labeled?
With deepfake fraud costing billions, is our financial system prepared for an era where seeing is no longer believing?