Chapter 11 · Field guide

How Can You Tell Bots and Coordinated Campaigns?

Short answer

You can rarely prove an account is a bot from the outside, but several signals together, such as huge posting volume from a new account, identical text across many accounts and synchronized timing, suggest automation or coordination worth treating with caution.

Key takeaways

  • No single signal proves anything: plenty of real people post a lot, use default avatars or have handles full of digits.
  • Coordination shows up across accounts: identical wording, the same links and posts landing within seconds of each other.
  • X offers a label for automated accounts that their operators can turn on, so labeled bots are the honest ones.
  • Calling someone a bot is a weak argument; judge the claim on its evidence instead.

Scroll through a busy thread and someone will call someone else a bot. Sometimes they are right. Often they are describing a real person who posts a lot, or simply disagrees with them. This chapter covers the general signals that point to automated or coordinated accounts, why each one is weak on its own, and what a skeptic does with a suspicion short of an accusation.

What signals suggest an account is automated?

Very high posting volume, a young account, repetitive content, round-the-clock activity and a thin profile are the classic signals of automation.

None of these is proof. Taken together, they shift the odds. Here are the common ones, with an invented example and an innocent explanation for each:

Signal Example Innocent explanation
Volume vs. age A three-week-old account with 9,000 posts A new, very enthusiastic user, or a news-alert account
Round-the-clock posting Posts every 20 minutes for 24 hours Scheduled posts from a small business
Repetitive text Same sentence posted 40 times with different hashtags A promoter who copies and pastes
Generic profile image Stock photo, cartoon or an oddly smooth face Many people protect their privacy with a non-photo avatar
Handle with many digits @name84726193 X suggests such handles at signup and many people keep them
Hashtag flooding Ten trending hashtags on an unrelated post Someone who misunderstands how hashtags work

Look at that last column before you decide anything. It is why the honest answer to "is this a bot?" is usually "maybe."

Why can't one signal prove anything?

Because real people produce every one of these signals, a single clue tells you very little, and stacking weak clues still leaves room for error.

Think about base rates, covered in misleading statistics. If a huge number of real users have digit-heavy handles, then "has digits in the handle" will flag far more humans than bots. The same goes for default avatars and heavy posting. A signal that fires on lots of innocent accounts makes a poor detector.

Researchers who study automation use dozens of features and still report error rates. You are looking at one profile on a phone. Hold your conclusion loosely.

What does coordination look like across accounts?

Coordination shows up as many accounts doing the same thing at nearly the same time: identical wording, the same links and replies landing within seconds of each other.

A single account tells you less than a group does. Patterns to watch for, using an invented example of a new snack brand:

  • Copy-paste replies. Dozens of accounts reply "Finally a snack that tastes like real fruit!" with the same typo.
  • Synchronized timing. The replies arrive within the same two minutes.
  • Shared links. Each one points to the same unfamiliar shop page.
  • Recycled profiles. The accounts were created the same week and share similar bios.

This is sometimes called astroturfing: activity built to look like grassroots enthusiasm. The broader term is coordinated inauthentic behavior. The accounts might be software, paid people or volunteers following a script. What matters for you is the effect. A claim can look popular without being popular.

What does X's automated account label mean?

X offers a label for automated accounts that their operators can turn on, which shows the account is automated and links it to the human-run account behind it.

According to the X Help Center, automated accounts perform programmed actions through the X API, and the label appears on the profile and may appear on posts. As of November 2026, X describes the label as still in testing. Many useful bots, such as weather or transit alert accounts, use it.

The catch is obvious: the accounts most likely to deceive you are the ones that will not label themselves. A label is good evidence of an honest bot. Its absence proves nothing.

What should you do instead of calling someone a bot?

Judge the claim on its evidence, avoid amplifying suspicious accounts, and use X's own reporting tools if you think rules are being broken.

"You're a bot" is an ad hominem in a new outfit. Even if it is true, it does not answer the claim, and if it is false you have insulted a real person. See logical fallacies.

A better routine:

  1. Check the claim, not the messenger. Would it hold up if a trusted source said it? Trace it to the original.
  2. Look at the source. The source credibility checklist walks you through account age, history and outside descriptions.
  3. Don't feed it. Replies and quote posts give coordinated accounts the visibility they want.
  4. Report through X. If accounts appear to break X's automation or spam rules, use the report option on the post or profile.
  5. Discount fake popularity. A thousand identical replies count as one voice.

Coordinated accounts work by faking a crowd. Your best defense is to ignore the size of the crowd and ask what the evidence says.

Frequently asked questions

Can I tell for sure if an account is a bot?

Usually not. From the outside you see only public signals, and real people can match many of them. Platforms have internal data you do not have. Treat your judgment as a level of caution rather than a verdict.

Are all bots bad?

No. Many automated accounts post weather alerts, earthquake reports, transit updates or reminders. X lets operators of automated accounts display a label that links the bot to the human-run account behind it.

What is coordinated inauthentic behavior?

It describes groups of accounts working together to mislead people about who they are or how much support an idea has. The accounts may be run by people, by software or both. The problem is the deception, whatever the tooling.

What should I do if I think accounts are coordinated?

Do not argue with them or quote them, because that spreads their posts. Use X's reporting tools if you think rules are broken, and judge the underlying claim on evidence from sources you can verify.

Sources

  1. X Help Center, About automated account labels
  2. X Help Center, Automation rules
  3. Vosoughi, Roy and Aral, The spread of true and false news online (Science, 2018)

Examples in this chapter are invented for teaching. Facts about platforms were checked in November 2026 and can change. See how we research.