Computational Propaganda

Signs it might be this

• Large numbers of accounts share very similar wording, timing and links, yet have no visible connection.
• Account histories are short, blank or inconsistent with the persona they present.
• Researchers or the platform itself have documented the network and removed it.
• The activity spikes around a political event and disappears afterwards.
• Accounts respond to criticism with the same stock replies rather than the content of your reply.

Signs it is probably something else

• The only evidence is that someone disagrees with you strongly and often.
• The accounts are real people who found the same post and shared it.
• You have judged by a single profile feature, such as a missing photo.
• Similar wording comes from a widely circulated news line.
• You cannot see the account’s history and are guessing from one post.
• The accounts are identifiable people or organizations who are open about their affiliations.
• Spread of the message tracks real events and real conversations.
• Independent researchers have looked and found ordinary activity.

Questions to ask yourself

• Does this account have a history that predates the topic, with signs of a real life?
• Is the pile-on built of accounts with nothing in common except the message?
• Would I judge the claim differently if I knew only a few people were behind it?
• Has a platform, researcher or journalist reported on this particular activity?
• Am I calling this fake because the evidence shows it, or because I cannot believe so many people disagree?

What it is

Computational Propaganda is political persuasion carried out with software. It uses automated accounts, networks of fake profiles and algorithm-boosted content that pass as ordinary people talking.

It differs from a single troll or a loud real user. It relies on volume and coordination that no group of individuals could produce by hand.

Parties and campaigns openly coordinate messaging and use scheduling tools, so automation alone is not the sign. Attribution is technical work, and a hard-arguing account is not evidence of a network.

Why it matters

The sense of what most people believe shapes what any one person feels able to say aloud. Here that sense gets set by accounts that are not people.

Real voices can be outnumbered, and honest users get wrongly accused of being bots.

How it shows up

• Political messages pushed through automated or fake accounts that imitate real, independent users.
• A matter of scale and speed. The same message can appear thousands of times in minutes, which a crowd of real people rarely does.
• Sometimes mixed with real users who are drawn in and share the content, which then spreads it further.
• Seen on screens as patterns in accounts and timing, rather than in what any single post says.
• Different from a hired commentator who writes under their own name, because the hallmark is simulated people.

What you can do

Pick what fits. You do not owe anyone a confrontation.

A check to run before you share or act

“Is there a real person behind this? Does the account have a life that predates this topic? If only a handful of people were posting this, would I still believe so many agree, or am I counting accounts instead of people?”

Ask about it gently

Before replying to a surprising account, check its profile and history. A single honest question such as ‘Can you tell me more about how you came to this view?’ shows whether anyone is there.

Set a clear limit

Decide not to share or boost content from accounts you cannot verify as real people, however well it fits your opinions.

Step back

Mute, block or leave a thread filled with identical replies, and avoid arguing with accounts that may not be people.

Get support

Report suspicious patterns, with links and dates, to the platform’s reporting tool, a fact-checking group or a reporter who covers online influence.

At work, school or with services

If you work in communications, moderation or research, keep dated records of what you saw. Follow your organization’s process for reporting suspected coordinated activity.

Using this label fairly

• Calling anyone who disagrees with you a ‘bot’ dismisses real people and ends conversations. The claim needs evidence about the account, and its opinions alone are not enough.
• Public accusations against a named account can wrongly expose an ordinary person to harassment.

How it differs from look-alikes

Astroturfing stages grassroots support by many means, including paid people and letters, while Computational Propaganda depends on software and fake accounts. Manufactured Consent shapes which real voices get coverage and needs no fake accounts.

Wedge Issue Exploitation is about which divisive topic gets pushed, and bots are only one possible amplifier.