LinkedIn Intensifies Its Fight Against AI Content

LinkedIn Intensifies Its Fight Against AI-Generated Content

LinkedIn, often described as the "Facebook of the business world," is ramping up its efforts to combat the influx of AI-generated content on its platform. In response to an increase in posts and comments produced by AI, LinkedIn has introduced new detection systems. These tools aim to limit low-quality posts, often referred to as "AI slop," a term that denotes content that, while polished in appearance, adds no real value.

A verification filter is also in development to facilitate the identification of bots and fake profiles. Laura Lorenzetti, Vice President and Executive Editor at LinkedIn Global Editorial, expressed in a blog post that excessive use of AI, especially at scale, dilutes the valuable insights that authentic human conversations can generate. She emphasizes that while AI can be a useful writing tool, contributions must reflect the user's personal voice and perspectives.

Microsoft and the Irony of the Situation

The irony of the situation is notable, especially coming from Microsoft, the owner of LinkedIn. LinkedIn's algorithms, like those of many social networks, tend to reward catchy and overly personal posts at the expense of those that actually contain substance. Moreover, Microsoft recently introduced a feature called Copilot in the browser, which encourages AI-assisted writing on the web. This feature was launched just a few days ago, with LinkedIn as a demonstration platform. This initiative seems to contradict LinkedIn's efforts to control the use of AI.

Detection Systems for More Authentic Content

To counter this phenomenon, LinkedIn is relying on technical systems developed with its editorial team. These systems are designed to detect content that appears to be AI-generated and lacks a clear perspective. The goal is to distinguish articles that provide context and expertise from those that are generic or repetitive.

Mass-produced comments generated by automation tools are also targeted. Responses that merely repeat the original post without substantial addition will be less visible in users' news feeds. Content flagged as AI-generated and lacking opinion will have reduced reach. Instead of appearing in the news feeds of other users, it will primarily remain within the author's network.

Initial Results and Upcoming Challenges

Initial tests from LinkedIn are promising, with a detection rate of generic content at 94%. Users are already reporting a decrease in unwanted posts. However, LinkedIn has not provided independently verifiable data, leaving some doubt about the frequency of false positives.

At the same time, LinkedIn is tackling AI-generated fake profiles that undermine authentic engagement. The platform is utilizing its verification system, with over 100 million members now verified.