Astroturfing is an organized campaign made to look like spontaneous public opinion. The name is a pun on AstroTurf, the artificial grass: fake grassroots. It ranges from a company posting glowing reviews under made-up names to governments running networks of accounts that flood social media.
Documented examples
- The 2017 net neutrality comments. When the Federal Communications Commission took public comments on repealing net neutrality rules, it received more than 20 million. A 2021 investigation by the New York Attorney General found that most were fake, including millions generated through lead-generation firms funded by a broadband industry campaign, and millions more submitted by a single person using made-up identities.
- State-run troll operations. U.S. Senate Intelligence Committee reports and the Mueller report documented how Russia’s Internet Research Agency ran fake American personas across Facebook, Instagram and Twitter around the 2016 election, on both the left and the right.
- Government “commentators.” A 2017 study by Harvard political scientists estimated that the Chinese government fabricates hundreds of millions of social media posts a year, mostly to distract and cheerlead rather than to argue.
- Platform takedowns. Meta, Google and others regularly publish reports on networks removed for “coordinated inauthentic behavior,” which describe the tactics in detail.
Claims about paid commenters were also a recurring theme on this domain’s former site; a widely cited 2013 story is in our archive.
Signs of a coordinated campaign
- Many accounts posting the same or nearly the same wording within a short time.
- New accounts, accounts with no history, or accounts that switch topics abruptly.
- Usernames that follow a pattern, such as a name followed by a long string of digits.
- Activity that rises and falls on a schedule that matches a workday in another time zone.
- Reviews or comments that praise or attack with unusual intensity but few specifics.
None of these signs proves anything on its own. Real people copy talking points and new accounts are often genuine. Coordination becomes clear when several signs appear together across many accounts.
AI makes this cheaper
Generative AI lowers the cost of producing varied, fluent text at scale, which makes copy-paste detection less useful. Platform researchers increasingly look at network behavior, such as who amplifies whom and when, rather than the words alone. See AI deepfakes and elections for the image and voice side of the problem.