What is a Twitterbot?
A Twitterbot is a software program designed to automatically follow Twitter users, like and retweet posts.
Often, the purpose of these bots is benign, such as driving engagement and providing helpful information when a certain keyword or hashtag triggers a bot response. Sometimes the purpose is commercial, such as driving post engagement and promoting goods and services.
But a bot's purpose can also be malicious, such as when they are designed to spread disinformation and propaganda through likes and retweets. This can create a false impression of a certain idea or news story being more truthful and widespread than it is.
Learn how to spot disinformation on social media here.
Although fake user accounts -- or hybrid human-AI accounts with automated elements -- can be found on every social media platform, they are especially prevalent on Twitter. A 2018 Pew Research Center study estimated that two-thirds of tweets containing links to popular websites were posted by bots.
The omnipresence of Twitterbots has even made Elon Musk waver on his offer to buy the platform. After making an offer to purchase Twitter, Musk said the deal could not move forward unless Twitter proved its claim that less than 5% of Twitter users represented spam accounts.
How are Twitterbots used?
Twitterbots -- like any social media element -- can be used to share information, encourage post engagement, and direct potential customers to a product or service. But they can also be malevolent, spreading misinformation; and creating spam messages, links and ads in an attempt to encourage click-through to legitimate business pages and sometimes phishing websites.
The most common and harmless uses of Twitterbots include the following:
- Automatically liking and retweeting posts. To help build a Twitter following and invite participation, a user might create or use a bot to like and/or retweet posts about specific interests or topics relevant to them and their community. For example, authors might use a Twitterbot to like and retweet links to reviews of their novels or bookstore sale announcements.
- Specifying who can reply to a tweet. By using the Twitter API, users can control who replies to a tweet. They can limit this to only those tagged or mentioned in a tweet or to followers of the user's account, which can help limit spam and other harassment.
- Generating automatic replies. A Twitterbot can be programmed to respond to a tweet or direct message once a trigger word, phrase or hashtag is detected. For instance, a Twitter user might post a tweet about being dissatisfied with a company's product. The bot could then send the user an automated reply or direct message letting them know they are being heard and that a customer service agent will reach out to them. It could also direct a user to the company's website for more assistance.
- Sharing important or interesting information. Many Twitterbots exist to provide users with helpful information, such as weather alerts. Others help to enhance user experience, such as the popular @threadreaderapp bot, which compiles threads into a more organized and easier-to-read format. Still other Twitterbots are simply for fun, such as Accidental Haiku (@accidental575), which retweets posts with the same syllabic structure as a haiku.
- Conducting polls and gathering market research data. A Twitterbot can be used to post polls on any topic, but they can be especially helpful for companies looking to gauge interest in a new product or service, raise awareness of an upcoming campaign, collect customer feedback or simply interact with existing and potential customers. Polls can also be used to drive traffic to a specific website.
How to spot a Twitterbot
For those who prefer a more organic and less automated Twitter experience, there are several ways to check if a user account is run by a real person.
Look for Twitter labels
In February 2022, Twitter launched a new feature that places a robot icon next to a Twitterbot's username and places an "Automated" tag under the username on tweets. Twitter has said this is part of its efforts to promote "good bots." While the labels are only used when the bot creator voluntarily enables them, a 2020 developer policy update informed developers all API-based bot accounts must clearly indicate the account is automated and state who runs it. Accounts not in compliance could be subject to suspension or deletion.
Check the suspected bot's post count
Because they are not run by humans who need to sleep, Twitterbot accounts often have abnormally high post counts. These accounts might post hundreds or even thousands of tweets per day, every day. Because it is highly unlikely a human would post that much, an account showing that much activity is almost certainly automated.
Check the suspected bot's post history
In addition to posting and retweeting frequently, Twitterbots tend to post about the same topic or handful of topics nonstop, sometimes repeatedly duplicating posts. Again, this type of activity is uncommon with human-run accounts.
Look for signs of humanity on the user's profile page
Twitterbots can sometimes be identified by suspicious profile pages that lack details a regular user would put in. If there is no bio or profile picture, or if the username is a random set of letters and numbers, the account could very well be a bot.
Look for original posts versus retweets
Bots, particularly those designed to post spam or disinformation, amplify other bots through retweets and quote links. If an account's post history contains nothing but those types of tweets, it is almost certainly a bot.
Use a bot-checking tool
Botometer is a website operated by the Observatory on Social Media and the Indiana University Network Science Institute. The site analyzes a given Twitter account's activity, along with its friends and followers, and gives it a score based on comparing the account to tens of thousands of example accounts. The higher the score, the more likely the account is automated.
Here are some common Twitterbot accounts and what they are used for:
- @earthquakeBot instantly tweets about any earthquake on Earth that registers as 5.0 or higher on the Richter Scale.
- @TinyCareBot sends hourly reminders to practice self-care. It tweets out reminders to drink water, look away from the computer screen, go outside, take a deep breath and more.
- @MoMARobot tweets four times a day, with each tweet containing a picture and information about a random piece in the Museum of Modern Art collection along with a link to the item's page on the MoMA website.
- @poem_exe selects random lines from existing works to create new poems, often with amusing results.
- @NiemanLabFuego finds and retweets popular stories among journalists.
- @colorize_bot uses AI to colorize any black-and-white image tweeted at it.
How to create a Twitterbot
For those interested in creating their own Twitterbot, Twitter has a full developer tutorial with code examples. In brief, you will need to complete the following steps:
- Create a Twitter developer account.
- Create an app through the developer homepage.
- Link the Twitter app and developer environment so the bot can communicate with the Twitter account.
- Program, test and deploy the bot.