This update continues our study of AI-written content on LinkedIn. We use the same nine topics, 100-word minimum and classification method to track how AI content changes over time.
In July 2026, we analyzed 5,000 public LinkedIn posts across nine topics. More than four in five (81.2%) were classified as Likely AI.
LinkedIn is also rolling out a native “Seems like AI slop” feedback option, giving members a direct way to flag posts and ads that appear AI-generated.
This is the same continuing study: nine topics, posts of at least 100 words, and Originality.ai’s 15% AI Allowance.
For July, we expanded the dataset to 5,000 posts. We collected 10 pages of results from 90 topic-and-date searches, then removed posts with the wrong date, under 100 words, or duplicates. We selected the final 5,000 before checking any AI scores.
The chart uses the full July dataset: 81.2% Likely AI across 5,000 posts.
This covers public LinkedIn posts found through our searches, not every LinkedIn post or anyone’s personal feed.
Our repeatable search method is built to track change over time. Feed-based studies measure what participating users see and can give more weight to popular posts. Our sample is narrower, but easier to compare consistently from month to month.
The earlier study ended with over half of long-form posts classified as Likely AI. In the much larger July analysis, 4,061 of 5,000 posts were Likely AI: 81.2%.
Likely AI content has moved from roughly half of sampled long-form posts in late 2024 to well over three-quarters today.


LinkedIn is rolling out a native “Seems like AI slop” option for posts and ads. It appears inside LinkedIn’s three-dot menu, beside “Not interested” and “Report post” or “Report ad.”


Sources and examples: 404 Media, LinkedIn’s feed policy, the Reddit rollout discussion, and posts from Navneet Kaur, Nicole Mezzasalma, Nick Bennett, Emmet Ó Briain and Matt Jones.
More than four in five LinkedIn posts in our larger July sample were Likely AI, continuing the upward trend.
LinkedIn’s new “Seems like AI slop” option shows that the platform is beginning to respond directly.
AI disclosure - The analysis was managed and reviewed by humans and significant AI editing on the post was used to help summarize the findings.

Using our proprietary Originality.ai AI detection tool, we analyzed DeepSeek Chat to find out if the LLM that disrupted the AI industry is detectable. Plus, see the latest AI detection accuracy for DeepSeek V3.2 (2025) and DeepSeek V4 Flash and V4 Pro (2026).