What Is “Foreign Inauthentic Activity”?

"Foreign inauthentic activity" is the term platforms, researchers and government agencies use for social media accounts, pages and groups that misrepresent who is behind them and where they operate, and that engage in another country's public debate under that false identity. The word "inauthentic" is doing the work. The problem is not that a foreigner has an opinion about an American issue. The problem is an account that claims to be a retired schoolteacher in Ohio when it is run from a desk in Dhaka, Shenzhen or St. Petersburg, and that uses the borrowed identity to make a message look like it comes from a neighbor.

Meta's formal category is "coordinated inauthentic behavior," defined as groups of pages or accounts working together to mislead people about who they are and what they are doing. X, TikTok and YouTube use similar language. The common elements are a false identity, a hidden operator, and coordination across more than one account.

Three kinds of operator

The activity comes from three distinct sources that get lumped together in headlines.

State operations. Governments or their intelligence services run networks directly. Russia's Storm-1516 network, which fabricates news stories on spoofed domains that imitate real outlets, is the best-documented current example. These operations are strategic, patient, and aimed at widening existing divisions rather than inventing new ones.

Contractors and proxies. Governments also hire private firms to run influence campaigns at arm's length. The China-linked networks that X and OpenAI disclosed in 2026 fall here. The tell is usually a mixed portfolio: the same personas that post about one domestic controversy also attack dissidents or defend the sponsor's foreign policy.

Commercial engagement farms. The largest layer by volume has no political sponsor at all. Operators in South Asia, West Africa and Southeast Asia build American-sounding personas and localized "Life in [State]" pages because American engagement is monetizable through Facebook's creator programs, affiliate links and page resale. These farms adopt whatever topic is generating reactions. In 2026 that topic is data centers. Before that it was inflation, immigration and school boards. Their output is often AI-generated imagery and text, which is why "AI slop" and "inauthentic activity" are increasingly discussed together.

How it works

The mechanics are consistent across all three types. Operators acquire aged accounts, either by buying them or by cultivating them for years with harmless content, because platforms trust old accounts more than new ones. They give the accounts names, photographs and biographical details matched to the target audience. They build audiences with nostalgic, sentimental or humorous content that older users share readily. Then they insert the message, in comments under news posts and advertisements, in group posts, or in the pages' own feeds.

The persona rarely survives close inspection. The friend list points to the operator's real geography rather than the claimed one. The biography is empty where a real person's would list a hometown, a school or a relative. The vanity URL, set when the account was created, often carries a different name from the one displayed. The gender field contradicts the photograph. A reverse image search on the profile picture returns a stock library or a stranger's LinkedIn page. Facebook's Page Transparency panel, where it is available, shows managers in a different country from the page's claimed home.

What the research has established, and what it has not

The independent research on the data center case is a good model for how to read claims about foreign influence generally. Graphika, Alethea and OpenAI have each documented real networks posting real anti-data-center content. Alethea's report is the most complete and traces all three operator types at once, including Facebook groups run from Bangladesh. Local reporters in Iowa and Missouri have identified specific pages and moderators with foreign friend networks.

The same research consistently finds that reach is small. Clemson's Darren Linvill checked the X accounts and found ninety posts that essentially nobody saw. Graphika's accounts drew little authentic engagement. Meanwhile Gallup finds 71 percent of Americans oppose a data center in their area, a figure no network of a few hundred accounts could produce. The defensible conclusion is that foreign inauthentic activity amplifies grievances that already exist. It does not create them, and treating a real local movement as foreign manipulation is itself a mistake the operators benefit from.