Propaganda Techniques and How to Spot Them

What Is Propaganda

Propaganda is the organized use of symbols, selected facts, and emotional appeal to move people toward a belief or an action. It works through framing and repetition rather than argument, which is why a piece of propaganda can be assembled entirely out of true information. What marks it as propaganda is the engineering behind it, not whether the content checks out.

The word began as a term for spreading doctrine. It comes from the Congregation for the Propagation of the Faith, a body of Catholic cardinals founded in 1622 to run missionary work. For many people it now carries a darker meaning tied to the atrocity stories and deceptive war aims of the two world wars and to the Nazis’ Ministry of Propaganda, Britannica notes.

Three related words get used interchangeably, and the differences matter for what you do about each one.

Propaganda is the widest of the three because falsehood is optional. A campaign ad that shows real job gains and hides real wage stagnation is propaganda. So is a speech that wraps a budget cut in flag imagery, and so is a genuine photograph posted with a caption from a different event. Misinformation and disinformation differ on what the person sharing a claim knew, the key step in a three-question test for misinformation and disinformation.

A Propaganda Operation on the Record

A 2018 federal indictment describes, in the government’s own words, what an organized propaganda operation looks like from the inside. On February 16, 2018, a grand jury in Washington charged the Internet Research Agency, a company in St. Petersburg, Russia, along with people and companies tied to it (indictment, United States v. Internet Research Agency LLC).

The indictment alleges that the organization “had a strategic goal to sow discord in the U.S. political system, including the 2016 U.S. presidential election.” It sought to conduct what it called “information warfare against the United States of America,” using invented American personas on social media. It ran like a business. By September 2016, according to the indictment, its monthly budget for the project “exceeded 73 million Russian rubles (over 1,250,000 U.S. dollars).”

The operation did not stay online. It ran a Facebook group called “Being Patriotic” and other fake American accounts, the indictment says. Through them it organized two rallies in New York in June and July 2016. On August 20, 2016, it held coordinated rallies across Florida, “posing as U.S. grassroots entities and U.S. persons.” It also paid Americans to take part, including one hired “to wear a costume portraying Clinton in a prison uniform at a rally in West Palm Beach.”

These are allegations in a charging document, not findings from a trial. What they show is the method: invented messengers, a steady budget, and real people moved to act by an organization that hid who it was.

The 12 Propaganda Techniques You Encounter Every Day

The Institute for Propaganda Analysis, which operated from 1937 to 1942, named seven core techniques. The table below adds five more that later researchers describe, including scapegoating, one of the techniques a 2022 Science Advances study taught people to spot. The original seven have held up because they describe moves rather than topics, so they work on a 1930s newsreel and a 2026 short video alike.

Three of the twelve work on emotion, which is why name-calling, glittering generalities, and fear appeals tend to land before you have read a source. Four exploit belonging, because bandwagon, testimonial, plain folks, and transfer all borrow credibility from someone or something you already trust. The last five bend the evidence itself.

Real propaganda runs several of these at once. A ten-minute rally speech can carry fear appeals, name-calling, bandwagon pressure, and card stacking, and the layering is what makes each one harder to catch, because you are busy reacting to one while the next goes by.

Why Your Brain Falls for Propaganda

The illusory truth effect

Repeating a statement makes people rate it as more likely to be true, even when they know better. A February 27, 2026 meta-analysis in Nature Communications pooled 182 studies and 366 effect sizes published from 1977 to 2025. Once the authors corrected for small-study effects, they called the effect “small but robust”. Earlier studies, they note, found it from age 5 to 80 and across trivia, news headlines, rumors, and marketing claims.

A small, reliable effect adds up at scale. Your brain treats familiarity as a shortcut for truth, so a claim you have met before feels easier to process and that ease gets misread as accuracy. Repetition is the central mechanism of a propaganda campaign.

Several obvious countermeasures fail. Paying people for accuracy does not erase the effect, and knowing the correct answer does not protect people either. In the meta-analysis, a general warning that some information might be false made no measurable difference, and earlier studies found a “disputed by fact-checkers” tag had little impact. Two things did help: asking people to judge whether a claim is true when they first see it weakened the effect, though only for about two days, and labeling a claim plainly as false reversed it.

The firehose of falsehood

RAND Corporation researchers Christopher Paul and Miriam Matthews used this term in a July 2016 paper to describe Russian state messaging, and later researchers have applied the model to domestic political communication. A firehose has four features. The volume is high, it runs across many channels at once, it is rapid and continuous, and it makes no commitment to truth or even to internal consistency.

A 2022 article in Frontiers in Political Science describes the effect: when leaders flood the zone this way, citizens retreat into cynicism and the belief that the truth is fundamentally unknowable. If the truth is unknowable, the authors write, reasoned debate becomes pointless because there are no agreed-upon facts.

Trump’s false claims accelerated every year of his first term, from about 6 a day in 2017 to 39 a day in 2020.

False or misleading claims per day

Period	Value
2017 (year 1)	6/day
2020 (year 4)	39/day
Change	6.5x

From Birtherism to the Big Lie

Birtherism shows how a conspiracy theory survives its own refutation. The factual question of where Obama was born had a documented answer in 2008, yet the claim kept circulating until Trump conceded the point in 2016.

The claim held a real share of the public. A CBS News and New York Times poll in April 2011 found that a quarter of Americans, and 45 percent of Republicans, thought Obama was not born in the United States.

Belief that the 2020 election was stolen has shown the same durability. PRRI found it changed little among Republicans between 2021 (66%) and 2023 (63%), even after the court rulings and recounts that rejected the claim.

How Propaganda Damages Democracy

Sustained distortion changes what a public is able to do together, which is a larger problem than any one person being wrong about any one fact. The damage tends to show up in eight ways. Democracy researchers count a shrinking space for free expression among the warning signs of authoritarianism.

Official framing of a group as a threat has preceded real harm. Executive Order 9066, signed February 19, 1942, cited “protection against espionage and against sabotage.” About 122,000 men, women, and children of Japanese ancestry were then forcibly moved to camps, and the government made no charges against them, the National Archives notes. In 1988, Congress acknowledged the injustice and apologized.

Researchers who study dehumanizing rhetoric describe a related risk. When public figures repeatedly cast a group as dangerous or less than human, they rarely give an order, yet the rhetoric can raise the odds that someone acts alone. That pattern is called stochastic terrorism, rhetoric that makes violence statistically likely without a direct command.

How AI Scaled Propaganda, 2023 to 2026

Generative AI cut the cost of running the twelve techniques rather than adding new ones. One measure of that drop is how many machine-run “news” sites now exist. NewsGuard’s first count, published May 1, 2023, found 49.

NewsGuard counted 49 AI-run “news” websites in April 2023 and 3,749 by June 2026.

AI-run 'news' websites grew from 49 to 3,749 in about three years

Category	Value
April 2023 (first count)	49
June 2026	3749

Sources: NewsGuard, Rise of the Newsbots, May 1, 2023; NewsGuard AI Tracking Center, updated June 23, 2026.

Those sites are aimed at more than human readers. NewsGuard’s March 6, 2025 audit of the Moscow-based Pravda network found that ten leading chatbots repeated its laundered claims 33% of the time, after the network published about 3.6 million articles in 2024. The tactic floods the open web so that a model trained or searching on it repeats the claims.

The tools people reach for to check a story often fail at it. In NewsGuard’s August 2025 audit, leading chatbots repeated false claims on news prompts more than a third of the time, nearly double the rate a year earlier, and its January 2026 audit still found a rate above 28%. AFP has documented chatbots confirming AI-generated images and video as real and passing off old footage as new strikes during India’s 2025 conflict with Pakistan.

People rarely catch good synthetic media. A February 2025 iProov study of 2,000 consumers in the U.K. and U.S. found that only 0.1% identified every real and fake image and video correctly, even though participants were told to look for deepfakes.

The institutional response has moved in two directions at once. Public Citizen’s tracker, updated September 29, 2026, lists enacted election deepfake laws in 32 states, one of them a Massachusetts provision that expired in February 2025. Courts have pushed back on the broader versions. On August 29, 2025, a federal judge in Kohls v. Bonta struck down California’s AB 2839 on First Amendment grounds, finding it could reach satire.

Where the Rules on AI Political Ads Stand

As of October 3, 2026

• A federal disclosure rule (H.R. 9922) Proposal

  Introduced July 23, 2026, by Rep. Joseph Morelle with 5 co-sponsors; referred to the House Administration Committee, no action since. (source)

• State election deepfake laws Law, 31 states

  Public Citizen’s tracker, updated September 29, 2026, lists enacted laws in 32 states, including a Massachusetts provision that expired in February 2025. (source)

• California's AB 2839 Blocked by a court

  A federal judge struck it down on August 29, 2025, in Kohls v. Bonta, on First Amendment grounds. (source)

Source: govinfo bill status; Public Citizen deepfake legislation tracker; court reporting

Verification capacity shrank over the same stretch. On January 7, 2025, Meta announced it would end its third-party fact-checking program in favor of a Community Notes system. The Washington Post’s Fact Checker, whose database produced the 30,573 count above, lost its longtime editor, Glenn Kessler, who left the Post in July 2025. Trust in the outlets that do this work fell decreased by ↓ 11 pts from March 2025 to Pew’s October 2025 report.

The money follows the attention. AdImpact projects $11.6 billion in political advertising for the 2026 cycle, more than the $11.2 billion spent in the 2024 presidential cycle. The World Economic Forum’s January 2026 Global Risks Report ranked misinformation and disinformation the second most severe global risk over the next two years, behind only geoeconomic confrontation.

How to Identify Propaganda

Propaganda is harder to catch than a plain lie because the individual facts often survive checking. These seven signals are about how a message is built rather than whether any one sentence is false.

1. Emotional intensity that arrives before evidence. If a message makes you angry or afraid before it gives you a source, treat that ordering as the signal. Feeling first and evidence second is the design.

2. The same phrasing across many outlets in one week. Organic ideas do not spread in lockstep, so identical wording on five channels means someone coordinated it. Ask who.

3. Vague positive language. Words that sound meaningful and define nothing, like “freedom,” “security,” or “real Americans,” are glittering generalities. Ask freedom from what, and security for whom.

4. One group blamed for a complex problem. Inflation, crime, and health costs all have several causes, so a message that assigns all of one to a single group is scapegoating.

5. Missing context around a number. A percentage with no base rate, a statistic with no source, a quote with no surrounding sentence. When a figure feels shocking, the useful question is what got cut.

6. Identity framing. “People like us believe X” is an argument about who you are rather than what is true, which is bandwagon and transfer working together.

7. Claims moving faster than anyone can check them. That speed is the firehose, and it is deliberate, because a claim you cannot verify in the moment still leaves a familiarity trace.

Recognition can be taught, and that is one of the few defenses with good evidence behind it. Researchers call the method prebunking, or psychological inoculation, and it works by showing people a weakened example of a manipulation technique before they meet the real thing. In a 2022 Science Advances study, five short videos, tested in six randomized trials and a field study on YouTube with 22,632 people, improved people’s ability to recognize manipulation techniques and to tell trustworthy content from untrustworthy content.

Where Persuasion Ends and Propaganda Begins

Propaganda is a specific category, not a synonym for messaging you dislike. Plenty of communication sets out to change your mind without meeting the definition.

• Advocacy is persuasion with its cards on the table. A nonprofit running ads for clean water legislation discloses its sponsor and makes claims you can check.
• Advertising is not automatically propaganda. A car ad that oversells the driving experience is marketing. It crosses over when it hides the sponsor, conceals the intent, or works through fear and identity rather than the product.
• Opinion journalism is not propaganda. An editorial arguing for a policy is an argument. It becomes propaganda when it poses as neutral reporting, buries contrary evidence, or substitutes emotional manipulation for reasoning.
• Repetition alone is not propaganda. Public health campaigns repeat “wear your seatbelt” because repetition works, and they do it with a named sponsor and a published evidence base.

The working test is three questions. Is the source transparent, is the evidence checkable, and is the intent disclosed? Three yeses describe persuasion. Any hidden answer moves the message toward propaganda.

The Honest Disagreement About Disinformation Research

Serious people disagree about what should be done, and both sides accept that propaganda works. The dispute is over who gets to decide what counts as false, and what a government may do once it has decided.

The case against organized counter-disinformation work comes from congressional Republicans and free-speech litigators. A November 2023 House Judiciary Committee staff report argued that the Election Integrity Partnership, a consortium led by the Stanford Internet Observatory, worked with federal agencies “to monitor and censor Americans’ online speech,” flagging posts for platforms. Subpoenas and lawsuits followed, and by June 2024 the Observatory had shed most of its staff, Tech Policy Press reported.

The case for that work comes from the researchers and their defenders. Paul M. Barrett of NYU’s Stern Center for Business and Human Rights called the Observatory’s collapse “a dangerous partisan victory” and “a serious loss in the campaign for rationality,” and described the congressional allegations against it as unsubstantiated.

The Supreme Court declined to settle it. In Murthy v. Missouri, decided June 26, 2024, the Court ruled 6 to 3 that Missouri, Louisiana, and five social media users lacked standing to seek an injunction against federal officials who had contacted platforms about COVID-19 and election content. Standing was the whole holding, so whether that contact crossed into unconstitutional coercion remains undecided.

The deepfake rulings give each side something to hold. California’s AB 2839 was struck down on First Amendment grounds even though the legislature had a real problem in front of it. That result tells the disclosure-first states how little a government may compel, and it tells the researchers that the constitutional limits on any fix are strict.

---

Frequently Asked Questions

What are the seven propaganda techniques?

The Institute for Propaganda Analysis, founded in 1937, named seven devices: name-calling, glittering generalities, transfer, testimonial, plain folks, card stacking, and bandwagon. They were collected in the 1939 book The Fine Art of Propaganda. Later researchers describe others, including scapegoating and emotionally manipulative language.

Is propaganda illegal in the United States?

Most of it is legal. The First Amendment protects political speech, including much speech that misleads, and the exceptions that fit propaganda are narrow, such as fraud, defamation, and incitement to imminent lawless action. The Smith-Mundt Act of 1948 set up the government’s information programs for foreign audiences; a 1972 amendment barred distributing that material inside the United States, and a 2012 law, in effect since 2013, lets it be made available here on request.

Is propaganda the same thing as public relations?

They overlap and are not identical. Britannica lists public relations among the fields where propaganda is used and notes that “public relations” is often used to avoid the word “propaganda.” Federal law requires paid political ads in federal campaigns to say who paid for them. By the test this page uses, the line is disclosure: a campaign that hides who is paying, or plants material designed to look like independent reporting, has moved into propaganda.

Is satire propaganda?

No, and courts have treated the difference as constitutionally important. Satire signals that it is not literal, and when a federal judge struck down California’s AB 2839 deepfake law in August 2025, the court found the law could reach satirical videos. Satire becomes a propaganda technique when the signal is removed and the piece is passed off as real footage or real reporting.

How do I check a claim quickly?

Find the original source before you find an opinion about it. Search the exact wording of the claim alongside the name of a wire service such as the Associated Press or Reuters, run a reverse image search on any photo, and check whether the outlet reporting it has a masthead and a corrections policy. Do not ask a chatbot to authenticate an image or a breaking-news clip. A NewsGuard researcher told AFP in June 2025 that chatbots are not reliable sources, particularly for breaking news, and AFP fact-checkers found Gemini confirming an AI-generated image as authentic.

What You Can Do

1. Learn the twelve techniques by name. Naming a move is what lets you see it, and the prebunking trials above found that teaching techniques improved people’s ability to recognize them. Start with the four you meet most often, which are name-calling, bandwagon, card stacking, and scapegoating.

2. Check the source before you share, especially on breaking news. Pew’s survey on fabricated news found that 23% of Americans had shared a made-up political story, and 16% had done it without realizing at the time. Open the original outlet, not the screenshot.

3. Ask who paid for the message. Find out who funded it, who benefits if you believe it, and whether the messenger profits from your reaction. Our dark money explainer covers how to trace funding through nonprofits that do not disclose donors.

4. Ask your state legislators for a K-12 media literacy requirement. Twenty-five states have a media literacy law on the books, per Media Literacy Now’s policy tracker, which covers laws passed through October 30, 2025. Look up your state’s status, then ask your state representative and senator by name whether they will introduce or co-sponsor a bill this session.

5. Pay for reporting that employs fact-checkers. Meta announced in January 2025 that it would end third-party fact-checking in the U.S., and the Washington Post’s Fact Checker lost its longtime editor that July. Subscriptions to local and independent outlets fund the people who do the verification that platforms stopped paying for.

6. Ask your representative to co-sponsor H.R. 9922. There is no federal rule requiring an AI-generated political ad to say so, which leaves 31 state laws of varying strength. The AI Transparency in Elections Act would require a disclaimer on ads substantially generated by AI. The letter below makes that ask.

If someone close to you is affected

Arguing harder rarely works, and PEN America’s guidance for families reaches the same conclusion the research does.

• Stay calm. Do not mock, shame, or corner them.
• Ask what source they trust and why.
• Focus on one claim at a time rather than the whole worldview.
• Lead with questions instead of lectures.
• Offer a credible source and ask if they will compare it with theirs.
• Keep the relationship open, because isolation makes the problem worse.
• Watch for signs the behavior is damaging their work, finances, relationships, or safety.
• If there is any talk of violence or self-harm, this is no longer a conversation problem. Call 911 if someone is in immediate danger. Call or text 988 (Suicide and Crisis Lifeline) if someone is in a mental health crisis, or text HOME to 741741 to reach the Crisis Text Line.

A useful opening: “I’m not trying to fight with you. I want to understand what you’re seeing and why it feels convincing. Can we look at one claim together?”

Organizations that track propaganda

• ADL tracks conspiracy theories and publishes guidance for families
• Brennan Center tracks election misinformation and state responses
• PEN America publishes scripts for talking with friends and family
• Reporters Without Borders runs a global Propaganda Monitor
• NewsGuard counts AI-run news sites and audits chatbot accuracy monthly