LinkedIn Fights "AI Slop" with New User Reporting and Advanced Detection
LinkedIn is taking a proactive stance against the proliferation of low-quality, AI-generated content, dubbed "AI slop," by introducing a new user-facing reporting feature. Users can now click a "Seems like AI slop" button when encountering posts they suspect are machine-written. This move reflects a broader industry concern, as platforms like Substack and even new startups grapple with an overwhelming influx of synthetic content and bot traffic, with internet infrastructure firm Cloudflare noting that bot traffic now surpasses human-generated requests. LinkedIn's Chief Product Officer, Hari Srinivasan, emphasized the company's commitment to preserving authentic human connection and expertise on the platform.
The new reporting button is just one component of LinkedIn's multi-faceted strategy to combat inauthentic content. The user feedback gathered from these reports will serve as a crucial signal, helping to train and refine LinkedIn's internal AI models to more accurately identify "slop." Complementing this, the platform is significantly investing in automated defenses, already blocking hundreds of thousands of automated comment attempts daily and millions of other automation efforts. Furthermore, LinkedIn is rolling out new AI classifiers specifically designed to detect AI-generated or low-quality content, particularly impacting what users see in suggested content recommendations from outside their immediate network.
For developers and researchers, LinkedIn's initiative highlights the escalating challenge of maintaining content integrity in the age of generative AI. The integration of user-driven reporting with sophisticated automated defenses and AI classifiers offers a valuable case study in building robust detection systems. This approach not only provides real-world data for improving AI detection models but also underscores the critical need for adaptive strategies to differentiate genuine human contributions from machine-generated noise. The platform's additional step of privately flagging potentially inauthentic content to creators, aiming to help them refine their AI usage, also points to a nuanced approach to content moderation that balances detection with user education.