Merriam-Webster defines astroturfing as organised activity intended to create a false impression of a widespread, spontaneously arising grassroots movement. [1]
The deception is not the message. It is the apparent number of independent people holding it.
What is actually being hidden
The lie in astroturfing is about origin, not identity. The people speaking may be entirely real, using their own names, saying things they have been paid or briefed to say — no fake accounts anywhere. What is concealed is whose interest is being served, and that concealment is the whole of the deception.
Astroturfing vs grassroots
The word is built out of the contrast with real grassroots organising. Merriam-Webster derives it from AstroTurf, which it glosses as metaphorically opposed to grass roots. [2] Artificial grass: something manufactured to look like the thing that grows on its own. A usage example in the same entry, from New Scientist, puts the practice as disguising an orchestrated campaign as a spontaneous upwelling of public opinion. [3]
So the opposite of astroturfing is not "a movement I agree with". It is a movement that arose without being organised into existence by a concealed interest — and that is a fact about origin, not about size, discipline, funding or how annoying you find it.
This is the part worth taking seriously, because the mistake is easy to make in the other direction. A real movement can be well organised, well funded and extremely effective. None of those is evidence of being manufactured, because none of them says anything about where it came from. "That's astroturf" is a convenient way to explain away a lot of people disagreeing with you without engaging with any of them — the mirror image of the tactic itself, and much easier to do by accident. If your only evidence is that a movement is too competent or too numerous to be real, you have no evidence.
The term's own history is a useful reminder that the real thing does exist. A usage example quoted in the dictionary's entry, again from New Scientist, reports that it appears to have been coined in 1985 by the then Texas senator Lloyd Bentsen, who noticed that the mountains of letters he was receiving about insurance legislation had originated with insurers. [4] Merriam-Webster's own entry dates the first known use of the word to 1993. [5] The two dates do not agree, and this site has no basis for resolving them — one is the dictionary's earliest recorded use, the other is a claim about who first said it.
Is astroturfing illegal?
That question has no general answer, and rather than give you a vague one: this site has nothing archived about the law in any jurisdiction, so it is not in a position to tell you. It depends on where you are and on precisely what was done. If it matters to you practically, it is a question for a lawyer where you live rather than for a reference page.
Astroturfing vs sockpuppetry
This is what separates it from sockpuppetry, which it is constantly confused with. Sockpuppetry is one person wearing several faces; astroturfing is a funded interest wearing the face of a spontaneous crowd. They frequently occur together, but neither requires the other, and mistaking one for the other sends you looking for the wrong evidence — fake accounts, when what matters is money and briefing.
The practical test is what would have to be true for you to be wrong. If the accounts turned out to belong to separate real people with ordinary histories, a sockpuppetry claim collapses and an astroturfing claim is untouched — because real people can be paid. That asymmetry is the reason the two need different evidence, and the reason a search for fake accounts can come back empty without telling you anything at all.
Why it works
People reasonably treat consensus as evidence. If many unconnected individuals have arrived at the same view, that is genuinely informative — which is exactly why manufacturing the appearance of it is worth someone's money.
It also shifts what looks normal. A position held by an apparent crowd is a position that can be stated without social cost, which is often the actual objective rather than persuading anyone specific. That is why the target is almost never the person arguing: it is the far larger number of people reading and quietly recalibrating what most people seem to think.
Recognising it
Signals, none of them proof on its own:
- Accounts with thin or recently-created histories converging on one topic
- Near-identical phrasing across supposedly independent voices
- Activity clustered into working hours in a particular timezone
- Enthusiasm that does not extend to any adjacent subject the account claims to care about
- No trace of the people involved existing anywhere else
Every one of these has an innocent explanation. New accounts are new for ordinary reasons; people who care about a topic do repeat each other's phrasing; campaigns do send supporters talking points openly. The pattern across many accounts is what carries weight, and even then it is consistent with coordination rather than evidence of it.
There is a harder problem underneath, and it is not about your judgement but about the shape of the thing. Researchers studying paid crowd-manipulation marketplaces found that many of the security techniques then protecting internet services assumed malicious activity was generated automatically by machines — and that such defences perform poorly or fail when real users can be organised to do the work by hand instead. [6]
The reason to put that on this page is a piece of this site's own reasoning rather than anything the study says: the list above is built on the same assumption. Thin histories, synchronised timing, identical phrasing — those are the fingerprints of accounts run cheaply at scale. Paid humans posting from their own established accounts leave none of them. The same authors report that their results suggest such campaigns are highly effective at reaching users. [7]
Two cautions about leaning on that. It is a single study, and it was published in 2012 — it describes what those marketplaces and those defences were then, and says nothing about what detection can do now, on either side. And it cuts both ways for a reader: it is a reason your failure to find bots proves very little, and equally a reason your impression of a crowd may be a crowd. What it does establish is that "these look like real people" is not the end of the question, which is precisely where the amateur analysis usually stops.
What to do
Very little, individually — and that is the honest answer. You are not the audience for astroturfing; the wider readership is. Arguing with the accounts is the least useful available action, because their number is the point and their individual arguments are not.
What helps is naming the pattern once, for the readership, and then leaving it. Platforms have reporting categories for coordinated inauthentic behaviour, and that is where the volume can actually be acted on. How to report, and what to expect from it.