AI-like writing is common in several recent social-media samples, but there is no reliable percentage for all posts across all platforms. The clearest estimates come from detector analyses of selected posts, and they measure likely AI involvement—not confirmed authorship.
What share of sampled posts looked AI-generated?
Originality.ai’s 2026 cross-platform comparison classified posts as likely to contain at least 15% AI-generated text. That threshold allows for mixed human-and-AI writing; it does not mean every word was generated by AI.
As an Amazon Associate I earn from qualifying purchases.
| Platform | Posts and collection window | Classified as likely to contain at least 15% AI-generated text |
|---|---|---|
| First seven months of 2026; the report also gives a separate July figure | 73.6% across the first seven months; 75.6% in July 2026 | |
| 195 randomly selected public, long-form English-language posts from July 2026 | 76% (126 posts) | |
| X | 201 English-language posts of at least 100 words from July 2026 | 63% (127 posts) |
| 559 posts of at least 100 words from June 2026 | 29% (162 posts) |
These are proprietary detector estimates, not a census. The samples differ in collection method, date, and—in the case of Facebook and X—specified language. Treat them as separate snapshots rather than a controlled ranking of platforms. Originality.ai also reported that 67.6% of its LinkedIn posts from July 2025 met its likely-AI classification, compared with 75.6% in July 2026. The company changed its collection method in November 2024 to include a newer public-search collection, a factor to keep in mind when interpreting trends. Originality.ai’s platform analysis.
The Tool Desk
Outbyte PC Repair FREEClear out junk files and repair common Windows errorsFree Scan →Outbyte Driver Updater FREEFix the driver behind crashes, sound loss and screen glitchesFind Drivers →A different estimate counts fully AI-generated writing
Pangram collected opt-in browser-extension scan statistics for 1,002,627 posts across LinkedIn, Medium, Substack, X/Twitter, and Reddit, beginning April 24, 2026. The extension scanned posts longer than 50 words. In that dataset, 25.72% of posts over 250 words were classified by its model as fully AI-generated. That is not the same measure as a post with at least 15% AI-generated text: mixed or AI-assisted writing is a separate category. Because users opted in to scanning, the data do not represent a probability sample of what everyone sees. Pangram also reported LinkedIn accounted for 62% of flagged AI content while making up about a third of scanned items; that describes the composition of its dataset, not LinkedIn’s share of all AI content online. Pangram’s analysis.
#1 Best Overall
Why there is no single answer for all social media
The studies measure different things. A detector may flag any likely AI contribution, estimate a minimum proportion of AI-generated text, or classify writing as fully AI-generated. Their samples also differ by platform, post length, language, date, and how posts were found. An opt-in scan, a public-search collection, and a random sample of public posts do not cover the same population.
Even a large percentage from one sample cannot establish the share of every post on a platform, much less across social media as a whole. Short posts, images, video, private accounts, other languages, and content outside the collection method may be missing or measured differently.
Detection is not proof of authorship
A detector score is an estimate based on patterns in text, not a record of how a particular post was written. Pew Research Center cautions that detection models can misclassify human-written documents as showing AI authorship and can miss AI-authored documents. Its July 2026 analysis of 490,000 English-language webpages sampled from Common Crawl found significant signs of AI authorship in 10% of the random page sample and in over one-third of pages published after ChatGPT’s public release. Those are web-page figures, not social-media prevalence estimates, and they do not verify individual authorship. Pew Research Center’s analysis.
Free tools Windows power users keep installed
One-click scans. No signup required.
Do users feel that AI slop is flooding their feeds?
In an online survey of 2,250 social-media users in the United States, the United Kingdom, and Australia, conducted February 5–9, 2026, 56% said they often or very often see “AI slop” in their feeds. Eighty-eight percent said AI-generated video tools had eroded their trust in social-media news. These results capture respondents’ perceptions in three countries; they do not measure the actual share of AI content or prove that AI videos caused a change in trust. “AI slop” is an informal label for low-quality AI-generated content, not a consistently defined research category. Sprout Social’s survey findings.
What the misinformation figures do—and do not—show
A European online-disinformation measurement summarized by MediaWell found AI-generated content in 24% of sampled TikTok misinformation posts and 19% of sampled YouTube misinformation posts. More than 83% of those AI-generated items carried no label. The denominator is misinformation posts in the study’s sample—not all posts on TikTok or YouTube—so these numbers cannot be used as platform-wide prevalence rates. MediaWell’s summary of the October 2025 measurement.
Independent reader supportYour contribution helps us test, update, and keep practical guides available for everyone.How platforms are responding, and what remains unclear
A 2026 CHI study examined AI-generated-content governance across 40 popular social platforms. Its abstract says just over two-thirds explicitly described such governance and identifies six kinds of action:
- Applying existing moderation policies to AI-generated content.
- Labeling or disclosure requirements.
- Restrictions specific to AI-generated content.
- Monetization constraints.
- Safeguards for AI-generation tools built into the platform.
- User resources or controls for managing feeds.
A stated policy does not establish how consistently it is enforced or whether it reduces exposure. The available estimates also do not settle how accurately detectors perform across social platforms, languages, formats, and post lengths. What they do show is that AI-like content appears frequently in multiple selected datasets—while the overall share of social-media content remains unknown. The 2026 CHI study.
Quick Recap
Best Value
Product prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon at the time of purchase will apply.
