Up to 15% of Accounts Are Bots on X

ByJames McCallef
x accounts: Editorial image of automated-looking social media accounts on screens

Bots on X are best estimated at roughly 9% to 15% of accounts, not a single precise count. The main official figure people cite, Twitter’s claim that fewer than 5% of monetizable daily active users, or mDAU, were spam or false accounts in Q1 2022, measured a narrower group than “all accounts” and came from the company itself.

That is why the answer feels slippery. Different sources measure all accounts, active monetizable users, or even share of tweet volume, which are not the same thing.

Current estimates for bots on X

The cleanest public summary from independent researchers is Reuters’ range of 9% to 15% of Twitter accounts being automated. If your question is “what percentage of accounts on X are bots,” that is the best-supported public answer.

Twitter’s own SEC disclosures said the company estimated false or spam accounts were fewer than 5% of mDAU in Q1 2022 and again fewer than 5% of mDAU in Q2 2022. But mDAU means users Twitter considered monetizable and active on a given day, not every account sitting on the platform. That is a narrower denominator, a small accounting choice with big consequences.

Older public estimates land in the same general neighborhood, but not exactly. A compiled historical overview says Italian researchers estimated about 10% of total accounts were bots in 2013, while a later academic estimate said up to 15% of Twitter users were automated accounts in 2017. Another figure repeated in public summaries put the total at 15% of all accounts, or about 48 million accounts, in 2020.

That 2020 figure also lets you do one useful back-of-the-envelope calculation: if 48 million accounts represented 15% of all accounts, that implies a total base of about 320 million accounts at the time. It does not tell you how many of those were active, monetizable, or posting every day, but it makes the scale concrete.

There is also a different metric that sounds bigger because it measures something different. An older Sysomos-based estimate found bots generated about 24% of tweets in 2009. That does not mean 24% of accounts were bots; it means automated accounts may have produced a disproportionate share of the content. A relatively modest share of accounts can still be very loud.

For the short version, the public evidence points to a range, not a magic number: Twitter said under 5% of mDAU; independent researchers summarized by Reuters said about 9% to 15% of accounts; older work suggested bots could drive around 24% of tweet volume. If you want the best general-purpose answer, it is the percentage of bots on X as a range of accounts, not the company’s narrower active-user metric.

Why the number varies

The first reason is simple: researchers are not measuring the same population. Twitter’s SEC filings referred to mDAU, Reuters’ summary referred to profiles or accounts, and older studies sometimes counted tweet output instead of account share.

The second reason is method. A 2023 arXiv paper on Twitter bot populations notes that estimates depend heavily on the bot-detection method and on the specific communities being studied. In plain English, a detector trained on spammy promo accounts may miss coordinated political bots, while a detector tuned to catch hyperactive posting may wrongly flag humans who just post a lot.

The third reason is that “bot” is not one thing. Some automated accounts are pure spam. Some are engagement-farming reply accounts. Some are scheduled news feeds, customer-service responders, or market-alert bots. Lumping them together can be useful for a headline, but it makes exact counting harder.

That is also why broad internet claims that bots surpassed humans in some contexts do not answer the narrower question of how many X accounts are bots. Platform bot share, human activity share, and web traffic automation are related ideas, but they are not interchangeable numbers.

A final caveat: the most widely cited official figure is old. The public SEC numbers here are from 2022, before the Twitter-to-X rebrand and before later product and moderation changes. There is no newer, equally authoritative public filing that settles the current count.

How to spot bots on X

No single sign proves an account is automated, but clusters of signs are useful. Public bot-detection research and platform investigations generally look for patterns like:

  • very high posting frequency across all hours
  • repetitive replies or identical phrasing
  • abnormal follower-to-following ratios
  • recent account creation paired with heavy activity
  • coordination with many similar accounts at once

A suspicious account often looks less like a person with odd opinions and more like a template with a pulse. It posts on a clock, reuses the same wording, and seems built to amplify rather than converse.

That said, spotting bots by eye is unreliable. The 2023 arXiv paper is useful here because it underlines the real problem: even formal detection systems vary depending on which communities they analyze and which behavioral signals they weigh. If researchers with classifiers disagree, random users definitely will.

The practical takeaway is blunt: there is no credible single public bot count for X today, but the best-supported estimate for accounts is still about 9% to 15%, while Twitter’s own narrower claim was under 5% of mDAU. If someone gives you one exact number with no denominator attached, they are probably selling certainty they do not have.

Key Takeaways

Frequently Asked Questions

what percentage of accounts on x are bots

The best public estimate is about 9% to 15% of accounts. Twitter’s own under-5% figure referred to mDAU, not all accounts, so it is not the same measure.

what percentage of users on x are bots

If by “users” you mean all accounts, the best-supported public answer is again roughly 9% to 15%. If you mean monetizable daily active users, Twitter said fewer than 5% in 2022.

how many accounts are bots on twitter

One public estimate put the number at about 48 million accounts in 2020, corresponding to 15% of all accounts. That figure is useful as scale, but it is older and not a current official count.

are bots on x the same as spam accounts

Not exactly. Twitter’s filings grouped “false or spam accounts”, while independent bot estimates often focus on automated behavior, whether or not it is classic spam. A scheduled weather-alert bot and a scam reply bot are both automated, but they are not the same kind of problem.

References

Further Reading

Last reviewed: 2026-06