LinkedIn Strengthens Its Fight Against AI-Generated Content

LinkedIn Faces the Rise of AI-Generated Content

LinkedIn, the professional social network, is facing a significant increase in content generated by artificial intelligence. A study conducted by Pangram, a company specializing in detecting AI-produced texts, revealed that nearly 40% of posts of 250 words or more on LinkedIn may not be written by humans. In light of this finding, LinkedIn has decided to take measures to counter this phenomenon, which it refers to as "AI slop," by introducing a new reporting button.

A New Tool to Report Suspicious Content

LinkedIn has recently introduced a button titled "Seems like AI slop," allowing users to report posts or comments they suspect have been generated by artificial intelligence. This initiative, announced by Hari Srinivasan, product director of LinkedIn Ecosystem, emphasizes the importance of moderation that relies on human intervention.

Two observations underpin this approach:

• On one hand, it is crucial to distinguish between the use of AI to refine ideas and the production of generic or inauthentic content.
• On the other hand, members should be able to receive feedback from real people regarding the authenticity of their posts, rather than relying solely on automated detectors that can be fallible.

Reports made through this button are intended to improve LinkedIn's detection models and optimize news feeds. The button is accessible by clicking on the three dots located at the top right of a post and then selecting the option "Seems like AI slop." Authors of reported content will be informed via their dashboard that some members consider their content inauthentic or overly influenced by AI.

LinkedIn Intensifies Its Fight Against AI Slop

For several months, LinkedIn has been working to reduce the dissemination of posts deemed too generic while avoiding complete removal. In addition to the new reporting button, LinkedIn has implemented several other measures:

• Strengthening Automated Defenses: LinkedIn claims to block hundreds of thousands of automated comment attempts daily. Over the past two months, billions of automation attempts, such as mass or low-quality posts, have been neutralized, according to Hari Srinivasan.

• Deployment of New Classifiers: These tools are capable of identifying posts that fall under AI slop or are of low quality, in order to limit their presence in suggested content and those from accounts outside the direct network of members.

• Replacement of the "Enhance Your Post" Feature: This feature, which reformulated texts using AI, has been replaced by a proofreading tool that corrects texts without altering the author's style.

• Extension of Profile and Page Verification: A new option also allows for blocking comments from company pages, thereby enhancing the security and authenticity of interactions on the platform.