Short answer
Bots tend to have default or stolen photos, handles with long number strings, recent creation dates, posting around the clock, generic or AI-sounding replies, and follower counts far out of line with engagement.
Key takeaways
- No single sign proves a bot; look for several together.
- X lets good bots label themselves as automated.
- AI-written reply spam is the main bot pattern of 2025-2026.
Bots range from helpful (earthquake alerts) to annoying (reply spam) to harmful (scams, coordinated manipulation). You can't always prove an account is automated, but the signs are usually there.
What do bot profiles look like?
Thin, generic and new.
| Clue | Why it matters |
|---|---|
| Default or stock photo, or an attractive stranger | Stolen or generated images are cheap |
| Handle like @maria84720193 | Auto-generated usernames |
| Joined very recently | Mass-created accounts |
| No bio, or a generic one | Nobody filled it in |
| Thousands following, few followers | Follow-spam behavior |
| High follower count, almost no engagement | Bought or fake followers |
How do bots behave?
They post constantly, repeat themselves, and reply fast with content that fits any post.
- Posting every few minutes, day and night.
- The same link or phrase across many replies.
- Replies that praise or rephrase the post without saying anything ("Great insight! This is so important.").
- Since 2024-2025, replies written by AI that sound fluent but generic, often with a link or a "check my bio".
- Instant replies to big accounts' posts, to sit high in the conversation.
What are good bots?
Openly automated accounts that say so, like weather, transit, news and art bots.
X lets developers label automated accounts with an "Automated" tag that names the managing account. If you're building one, see how to make an X bot.
What should I do about bad bots?
Don't engage, then mute, block or report as spam.
Replying feeds them attention. Report clear spam, especially scams using a famous name. If bots follow you, remove them or ignore them; they don't affect your reach much.
Can a checkmark or a big following prove an account is human?
No; both can be bought.
Paying for X Premium gives a blue check to anyone who passes the basic review, and followers can be bought in bulk. Better signals of a real person: a long history of varied posts, conversations with people who reply back, photos and details consistent over years, and links to other profiles that check out. Even then, stay skeptical of accounts that suddenly change topic or language, which can mean the account was sold or hacked.
Questions people ask
Are all bots bad?
No. Weather alerts, earthquake feeds, transit updates and art bots are useful. X asks automated accounts to add an 'Automated' label naming the human who runs them.
How do I report a bot?
Use Report on the account or post and choose spam. For networks of fake accounts, report several examples.
Can a checkmark account be a bot?
Yes. A blue checkmark means a Premium subscription, which spam operators also buy. It isn't proof of a human.