October 2026
LinkedIn Cracks Down on AI Slop, YouTube Changes How Views Count, Threads Steps Out on Its Own
If your LinkedIn feed has felt increasingly AI-generated lately, you're not alone. In Q3, platforms began pushing back against low-quality AI content while embedding AI more deeply into search, discovery, and content creation. LinkedIn also continued refining what earns reach, YouTube changed how it counts views, and Threads took another step toward becoming a standalone platform. A few of the updates most relevant to corporate communicators are below:

- LinkedIn cracked down on “AI slop”, introducing a way for users to flag low-substance, low-effort content that appears to be AI-generated, significantly decreasing the views on flagged content
- LinkedIn says it isn't penalizing AI-assisted writing broadly; the focus is on generic content lacking original experience, perspective, or insight
- LinkedIn adjusted its feed display to favor comments and replies that are relevant to individual users, rather than treating all engagement equally
- LinkedIn is expanding verification tools, including ways to validate employment history and allowing companies to remove people who make false claims
- LinkedIn is testing a Network tab consisting of posts from people and companies a user directly follows or is connected to, rather than algorithmically recommended content

- X expanded tools designed to help users understand when posts or accounts are subject to distribution restrictions, including additional information about potential reach penalties
- X's latest disclosures under the EU's Digital Services Act showed another decline in its European user base

- Starting in September, businesses no longer need a matching Instagram identity or account to advertise on Threads

- YouTube standardized public view counts across Shorts, long-form videos, podcasts, and livestreams, so a public view now registers as soon as a video begins playing instead of waiting for a viewer to actively engage with the content; the previous measurement remains available in Analytics as Engaged Views; YouTube explicitly warned that public view counts may increase faster under the new methodology
- YouTube rolled out Ask YouTube, a conversational search experience that lets viewers ask questions and discover content using natural-language prompts
- YouTube is expanding its testing tools beyond titles and thumbnails, with a new feature that will allow creators to test up to three different cuts of a video to determine which version holds audience attention best

- Reddit reported more than 130 million daily active users in Q3, continuing its steady audience growth.