Social Media AI Tracker Report: August 2026

Social media companies are undertaking efforts to reduce AI-generated content on their platforms, and those efforts appear to be making a modest impact. 

Here’s what we saw for August:

Facebook and X each saw 6% decreases in Likely AI content in August, while LinkedIn experienced a reprieve from an otherwise relentless increase.

Reddit, meanwhile, saw a 2% uptick in Likely AI posts (from June to July 2026). However, since our Reddit data is on a one-month delay due to data availability constraints, it too may have seen a decrease in August. 

While Reddit saw the only month-over-month increase in Likely AI content, the platform features nearly half as much Likely AI content as the next in our ranking (X formerly Twitter) and remains the only one well below the 50% threshold. 

The slowdown among most major social media platforms may be the result of newly rolled out or beefed-up features designed to reduce or discourage AI use. 

For example, August was the first month that LinkedIn users were encouraged to report when a post “seems like AI slop.” In recent months, X/Twitter and Meta have also experimented with new policies and approaches designed to discourage AI-generated content and encourage original content on their platforms. 

While those efforts appeared to have made a modest impact, the change is relatively incidental, given the significant rise in Likely AI content across all social platforms in 2026 overall. 

The question that remains is whether the decreases we saw in August were the beginning of a trend towards less AI use on social media, or a temporary slowdown in the otherwise relentless explosion of AI content on social platforms.

Threads: New Addition for August

Our August report is also the first to feature Threads, which will be included in all future reports. The short-form text-based platform enters our monthly ranking in second place behind LinkedIn, with 66% of posts determined to be at least 15% Likely AI. 

Key Findings

• Most sampled posts on every social media platform other than Reddit appear to be at least 15% AI-generated.
• Efforts to reduce AI content on platforms like LinkedIn, Facebook, and X appear to be making a modest impact but aren’t enough to meaningfully counter the meteoric rise seen this year.
• Despite some recent reductions, AI content is up significantly across all major platforms over the last year.
• Shorter LinkedIn and Facebook posts are more likely to leverage AI, but the opposite is true on X and Reddit.
• On LinkedIn, 87.5% of posts related to leadership or data and analytics were Likely AI. As were 85% of crypto- and finance-related posts on X. Nonprofits and civic Facebook groups also saw an equally high proportion of likely AI content, at 85%.
• LinkedIn remains the biggest hotbed for AI-generated content. However, more than half of written posts on most social media platforms appear to be at least 15% AI-generated.  

Social Media: Cross-Platform Comparison

The data below represents how much each platform’s long-form written content is more likely than not to be at least 15% AI-generated.

• LinkedIn: 76% of sampled posts Likely AI (unchanged)
• Threads: 66% of sampled posts Likely AI (new for August)
• Facebook: 59% of sampled posts Likely AI (down 6%)
• X (formerly Twitter): 57% of sampled posts Likely AI (down 6%)
• Reddit: 31% of sampled posts Likely AI (up 2%)

LinkedIn – 75.6% Likely AI

Originality.ai analyzed 119 randomly selected long-form English-language LinkedIn posts published between August 4 and 31 of 2026. The analysis determined that 90, or 75.6%, were Likely AI (at 15% AI Allowance settings).

LinkedIn continued its reign as the most AI-heavy social media platform in August, with 76% (or unrounded 75.6% of posts classified as Likely AI), despite the platform’s efforts to discourage AI usage through the introduction of its “Seems Like AI Slop” reporting button on July 30. Based on these findings, the feature does not appear to be making an impact, though recent updates may prove more effective in our next report. 

Though Likely AI content on LinkedIn was static between July and August, other platforms widened that gap following modest reductions. 

Shorter posts continued to feature more AI content, with one exception. 

Posts of 150-249 words saw a slight reduction in Likely AI content between July and August, from 75% to 73%, while posts of all other lengths saw increases. Those of 100 to 149 words, for example, saw a significant jump from 69% to 86% Likely AI content. Only half of posts 400 words or longer, meanwhile, were more likely than not at least 15% AI-generated.

Posts related to leadership and data and analytics continued to see the highest proportion of likely AI content at 87.5% each, followed by digital finance at 83%, and both climate and emerging technology at 73%. AI-related posts, ironically, were somewhat less likely to be AI-generated, at 70%, followed by careers content, which saw only about a third of posts flagged as likely AI. 

Threads — 66% Likely AI

Originality.ai analyzed 162 randomly selected English-language Threads posts of 50-99 words published in August and determined that 107, or 66%, were more likely than not at least 15% AI-generated. (There will be more robust data on Threads in future reports). 

Facebook — 59% Likely AI

Originality analyzed 166 randomly selected long-form English-language Facebook posts of 100 words or more published in August of 2026 and determined that 98, or 59%, were Likely AI.

The finding represents a 6% month-over-month decrease in likely AI content on Facebook, down from 65% in July, and a significant reversal in the rapid rise of Likely AI content on Facebook, which has trended sharply upwards through 2026. 

At the start of the year, our detector determined that 49% of Facebook posts were Likely AI-generated, and that number continued to rise each successive month until reaching a peak in July, before the first decline of the year in August. This month’s 6% drop puts the proportion of likely AI content on Facebook back to levels last seen in May. 

Shorter Facebook posts remained more likely to leverage AI content in August, with 62% of those in the 100-to-199-word range likely AI-generated, compared with 54% of posts of 200 to 299 words and 42% of posts 300 words and longer.

Within Facebook groups, nonprofits and civic groups continue to see a significant proportion of Likely AI content, with 85% of posts flagged as Likely AI, up from 82% in July. 

Second was public figures, education, and creators at 69%, up from just 50.9% in July, followed by health-related groups at 63%. While culture, museums and libraries, and higher education categories ranked lowest for Likely AI content, at least 50% of posts in every Facebook group category were likely AI-generated.

X (formerly Twitter) — 57%

Originality.ai analyzed 174 randomly selected long-form English-language X posts of 100 words or more published in August of 2026 and determined that 99, or 56.9%, were likely at least 15% AI-generated.

The latest finding represents a 6% month-over-month drop in likely AI content on X as compared with July’s 63%. Though we can only speculate what caused the decline, the platform appears to be making attempts to reduce AI-written posts, according to a February 2026 post by Nikita Bier, who was head of product at X at the time (and has since stepped down from the role). 

Unlike most platforms, X has seen a modest decline in likely AI content this year following its December 2025 peak of 68%. Likely AI content on the platform has now reached its lowest point since August of 2025, when 50% of posts were determined to be likely AI-generated. 

Unlike Facebook and LinkedIn, longer X posts typically feature more AI content, with just 52.4% of posts of 100 to 149 words likely AI-generated, compared to 54% of posts 150 to 199 words and more than two-thirds of posts 200 words or longer.

The finance and crypto category remains the biggest hotbed for likely AI content, with a whopping 85% of posts deemed likely AI-written, followed by technology, AI and science at 75%, health and wellness at 71%, and business, work and marketing at 65%.

Politics, news and public affairs-related posts on X were 59% likely AI-generated, while personal life and general commentary trailed at 52%.

In the month of August, only faith and spirituality (40%), sports, entertainment and gaming (38.5%), and education and self-improvement (25%) posts remain below the 50% likely AI threshold.

Reddit — 31% Likely AI

Originality.ai analyzed 564 randomly selected long-form English-language Reddit posts of 100 words or more published in July of 2026 and determined that 173, or 30.7%, were likely to be at least 15% AI-generated.

That represents a modest 1.7% increase from our previous report, which found 29% of posts were likely AI-generated in June. It is worth noting that our Reddit reading is on a one-month delay due to data collection limitations.

As with X, longer Reddit posts tend to feature more AI content, especially posts of 700 words or more. Just a quarter of posts of 100 to 199 words and less than a third of those in the 200-to-799-word range were likely AI-generated, compared with nearly two-thirds (65%) of posts of 800 words or more.

This finding may suggest that comments, which tend to be shorter, are less likely to be AI-generated while posts, which tend to be longer, may leverage more AI assistance. Likely AI usage in longer posts has also increased each month of this year as compared with the same month last year.  

Methodology

While we previously only included posts of at least 100 words in our rankings to increase the detector's accuracy, we adjusted that standard to posts of 50 to 99 words for Threads, given the short-form nature of the platform. As a text-based AI detector, our analysis is also limited to written content. 

A post is deemed Likely AI if the tracker is more confident than not (50% confidence or higher) that at least 15% was written by AI.

It is also worth noting that the tracker does not necessarily represent the proportion of AI-generated content a given user will see while browsing their feeds or timelines, as each user experience differs based on the platform’s algorithms and filters. That means some users are served more AI-generated content than others, even on the same platform.