Blog
What are bots on social media, and how do you spot them in your comments?
A bot on social media is an account or a program that posts, comments, follows or replies automatically. There are two very different kinds: the chatbot a brand connects through the official route to answer people who message it, which Meta documents and allows under conditions, and automated or fake accounts that inflate engagement or leave spam, which Meta’s and TikTok’s rules prohibit. In your brand’s comments, the second kind gives itself away through patterns (generic text, links, offers, bursts of look-alike accounts) rather than any single sign, and the right move is to filter or hide those comments without replying.
We read every source on this page on November 19, 2026. Platforms change their rules, so check each link before you act on it.
An example: five comments under the same post
A fictional case, to practice reading them. A café posts its holiday opening hours on Instagram and gets these comments in the first hour:
| Comment | What you can see | What to do |
|---|---|---|
| “Get 1,000 real followers in 24 hours, link in my bio” | Offers engagement in exchange for something | Hide or delete. Don’t reply |
| “I’m from Instagram support, your account will be suspended, verify it here” plus a link | Claims to be the platform and pushes a link | Don’t open the link. Delete |
| “Great content, check your DM”, posted word for word by six accounts with no profile photo within two minutes | Same text, similar accounts, all at once | Hide, then look at those accounts |
| “Are you open on the afternoon of the 31st?” | A question about the post | Reply |
| “Looks delicious” from an account with years of photos of the city | Generic, but from an account with a history | Reply or leave it |
A person wrote the fourth and fifth comments, even though the fifth is as short as a bot comment. What sets the first three apart is what they ask for (a click, a message, a purchase) and the fact that they repeat. That’s the reading that works on your account: what the comment wants, and whether it shows up the same way elsewhere.
What counts as a bot, and what doesn’t
People use “bot” for two things the platforms treat in opposite ways.
The authorized bot. This is automation that connects through the official API and works within its rules. The Instagram messaging API lets an app send and receive messages between an Instagram professional account and its customers, with three conditions Meta spells out in that documentation:
- The app can only message a person after that person has messaged the account.
- It has to reply within the messaging window, which for most user actions lasts 24 hours.
- When applicable law requires it, the automated chat must disclose that the person is talking to an automated service: at the start of a conversation, after a significant lapse of time, or when the chat moves from a person to automation. Meta recommends doing it anyway, as best practice, even where the law doesn’t require it.
The bot the rules prohibit. Neither Meta’s spam policy nor its inauthentic behavior policy uses the word “bot.” They describe behavior:
- Posting, sharing, engaging with content or creating accounts at very high frequencies, whether manually or automatically. Meta adds that it may restrict accounts acting at lower frequencies when there are other signs of spam, such as repetitive content.
- Selling, buying or exchanging engagement, such as likes, views, follows or clicks.
- Sharing deceptive links: ones that promise one kind of content and deliver another, or landing pages that pretend to be a well-known brand.
- Using inauthentic accounts, Pages or other assets to deceive people about the identity or origin of whoever is behind them.
TikTok’s integrity and authenticity guidelines, in effect since September 24, 2026, do name automation: “We strictly prohibit automation tools, scripts, or other tricks designed to bypass our systems.” They also prohibit the trade of services that artificially boost engagement, and say that when TikTok detects inauthentic metrics it removes fake likes, followers and other inflated signals.
Which apps you can connect to Instagram without putting the account at risk, and how to recognize one that asks for your password, we cover in what it means when Instagram suspects automated behavior.
Chatbots that answer a brand’s messages through the API are bots of the first kind, the authorized one. If you’re evaluating one, it helps to tell it apart from an AI agent, which decides the steps and uses tools such as your calendar or your order system: we explain the difference in AI agent vs chatbot.
Are social media bots illegal?
This page doesn’t answer that as a legal question: it depends on the country, the use and who’s behind it, and for your own case you should ask a lawyer. What the sources we read do put in writing are the platform rules above, and Meta’s own documentation ties the automation disclosure to what applicable law requires, singling out users and markets in California and Germany. If your concern is buying followers or engagement, we go through the US FTC rule on fake indicators of social media influence in what happens if you buy followers.
How to spot a bot comment on your account
We didn’t find a primary source that measures what share of social media accounts or comments are bots, and the policies we read don’t publish one, so we don’t give a percentage. On your own account these signs are more useful. We propose them based on the behavior Meta prohibits. None proves anything alone; they matter when they stack up:
- It asks you to leave the post. A link, “check your DM,” “message me privately” or “link in my bio” in a comment that has nothing to do with what you posted.
- It offers something for engagement. Followers, likes, views or prizes. That’s exactly what Meta’s spam policy prohibits selling or exchanging.
- It speaks for the platform or another brand. A suspension notice, “support,” or a giveaway you didn’t run, almost always with a link.
- It repeats. The same text, or nearly the same, across several of your posts or from several accounts.
- It arrives in a burst. Lots of similar accounts commenting within a few minutes, with no connection to your usual audience.
- The account has no history. No photo, no posts, a name made of letters and numbers. This is the weakest sign: a real person who never posts looks the same. To settle it, on Instagram you can open About this account: Instagram’s help says it shows the date the account joined and how many times its username has changed, and that this count may help identify whether an account is authentic or misleading.
A short, friendly comment with no link and no request rarely calls for more than a quick reply or leaving it alone. If the suspicious accounts are among your followers, the possible spam list and how to review it are in the same guide on buying followers.
What to do with bot comments
- Don't reply, and don't open the link
Your reply changes nothing for the bot, and a reply from your brand under a scam can read as an endorsement. The link can lead to what Meta’s spam policy describes: deceptive redirects and pages that imitate a well-known brand.
- Turn on the platform's filters
On Instagram, the help page on hiding comments describes Hide unwanted comments, which is on by default and hides comments with common offensive words, phrases or emojis, and Hide spam. Under Hidden Words you can add your own list with the phrases that keep repeating, such as “check your DM.” On YouTube, the help page on comment settings lets you hold potentially inappropriate comments (at the Basic or Strict level) and comments that contain links; they wait in the Held tab in YouTube Studio for up to 60 days and aren’t published unless you approve them. Comments from your approved users skip those filters. If you also moderate TikTok, see how to turn on TikTok’s comment filters.
- Hide before you delete
On Instagram you can only hide comments manually on your own posts, from the iOS or Android app, by tapping Hide under the comment. According to the help page, a manually hidden comment is visible only to you and the person who wrote it, and that person can’t see that you hid it. If you got it wrong, tap View hidden comments at the bottom of the comments and unhide it. Save deleting for obvious spam and scams.
- Check the first few days
Look at what the filters caught and pull out anything a real person wrote. A word that’s too broad on your list will hide customer questions too.
If you run several accounts, HeyMark moderates automatically: it hides, holds for review or deletes comments, with your approval to delete. How it’s set up per brand, and what happens to each flagged comment, is in social media comment moderation with AI.
Plan your next post in HeyMark.
Keep the idea, review the draft with your team, and see how it performed in the accounts you connected.