Who Gains If You Believe It?
Stephen Petro runs a YouTube channel about critical thinking and teaching yourself. On November 16, 2026, he posted a video about 14 minutes long called This Curriculum Quietly Vanished From Schools. The thing that vanished is a one-page guide from 1937 called “Some ABCs of Propaganda Analysis.” It is seven short instructions, one for each letter from A to G.
I build software with language models, and I write this blog. Every day I read claims about what AI is about to do to work, to schools, and to the country. Most of those claims come from someone who wants something from me. I wanted a checklist that came from somewhere else. That is what pulled me into this video.
One note before we start. The video is my starting point, not my only source. Everything about the 1937 guide that goes beyond what Petro says comes from two scholarly papers and one report, and I name them at the end.
Why 1937, and why now
Petro’s argument for the guide is that it is old. A book about persuasion written this year carries the writer’s feelings about social media, AI, and politics. A guide written in 1937 carries none of that.
Propaganda, as the Institute used the word, is “expression of opinion or action by individuals or groups deliberately designed to influence opinions or actions of other individuals or groups with reference to predetermined ends.”
That definition is broader than most people expect. Under it, propaganda does not have to be a lie, and it does not have to come from a government. An ad, a campaign speech, a charity appeal, a lesson plan, and a YouTube video all count. The video puts itself in that list.
The group behind the guide was the Institute for Propaganda Analysis, or IPA. The story starts on March 29, 1937, when Miller met the department store owner and philanthropist Edward Filene at a club in Boston. Filene worried that Americans were losing the ability to weigh information, and he asked for ideas. Miller drafted a proposal. The group met again in New York, where Filene offered Miller a $10,000 grant on the spot. By that fall they had settled on three years of money for an institute at Columbia Teachers College, and they named it the Institute for Propaganda Analysis.
The man who sent a man to prison
The first instruction makes more sense once you know who wrote it.
Miller started as a reporter for the Cleveland Plain Dealer in 1916, and his beat was the federal prosecutor’s office. In June 1918 he was sent to Canton, Ohio to interview the socialist leader Eugene V. Debs. Debs gave a speech against the war to a crowd of more than 1,200, with a government stenographer in the room. Miller took notes in shorthand, filed his story, and called the prosecutor to say Debs had broken the law. He later became the lead witness for the prosecution. Debs was sentenced to ten years in prison.
The Espionage Act of 1917, tightened by the Sedition Act of 1918, made it a crime to speak against the war effort. That is the law Debs was convicted under.
All of this comes from Miller’s own account, published in The Progressive in 1963. That matters, because it is a man describing his own part in jailing a famous socialist, forty-five years later.
Miller’s 1963 account adds one detail the video does not include. He wrote that word came back from Washington saying they did not want to prosecute Debs. Miller and his paper kept up the pressure until the prosecutor went ahead.
Then Miller spent time in France, and his view of the war changed. In his own words, “In six months in France, I learned more about how wars are made than in all my years in college. I learned how propaganda worked. I decided Debs was right and I was wrong.”
Before Debs went to jail, he ran into Miller and proposed a deal. If they both made it back, they should meet and “tell each other who was more nearly right or wrong about this war.” Miller later pressed for Debs’s release. President Harding let him out in time for Christmas 1921.
So the man who wrote the guide started from the fact that it had worked on him.
The seven letters
Here is the guide. The 1937 wording sits in quotation marks.
A. “Ascertain the conflict element in the propaganda you are analyzing.”
Before you decide whether a message is true, find the fight underneath it. Who is in conflict, what does each side want, and what does each side gain if you believe its version?
Petro’s worked example is a sentence I read almost daily: AI is coming for your job, and if you do not adapt now you will be left behind. On its face that is a conflict between workers and machines. Look closer and there are more sides. Workers want security. Employers want more output at lower cost. Technology companies want fast adoption. Politicians want to look either protective or pro-innovation. And people selling AI courses do better when the future feels scary and urgent.
None of that makes the warning false. It means the warning comes out of a fight, and the people in that fight want different things from you.
B. “Behold your own reaction to this conflict element.”
What did the message make you feel? Anger, relief, fear, the warm glow of being right?
Petro uses his own title as the example. If you already believe schools have gone soft, the title feels good. It tells you something valuable was taken, and that you are one of the few smart enough to go get it back. If you work in public education, the same title reads as dishonest. Neither reaction says anything about whether the 1937 guide was real. Both reactions change what you notice next.
C. “Concern yourself with today’s propagandas associated with today’s conflicts.”
The guide warned against getting very good at spotting propaganda that can no longer touch you. It is easy to look at a wartime poster and feel clever. The hard version is the message attached to something you are invested in right now: your job, your health, your income, your religion, your party.
The rule the video draws from this: pair every old example with a present one. If you can name the emotional tricks in a politician you hate but never in a creator you trust, “including me,” you are studying propaganda as entertainment.
D. “Doubt that your opinions are ‘your very own.’“
The guide says plainly that your beliefs were shaped by your family, your neighborhood, your religion, your income, your job, your party, your friends, and the part of the country you grew up in. That does not make you incapable of thinking. It means your surroundings set which claims look obvious, which sources feel safe, and which possibilities never come to mind at all.
Petro suggests a test. Pick one belief you hold hard. When did you start believing it? Who around you believed it? What would it cost you socially to drop it? If you had been born into a different family, class, religion, or party, how likely is it you would have landed on this same conclusion?
A belief can have a social cause and still be true. But if you never checked why you believe it, you do not know whether it holds up or just got lucky.
E. “Evaluate, therefore, with the greatest care, your own propagandas.”
This is the letter that stings. It is not enough to admit your surroundings shaped you. You also have to look at the selective stories, loaded words, convenient gaps, and heroic versions used by your own side.
Petro goes back to the AI example. If you already believe AI will replace most people in your field, do you treat every layoff as confirmation and skip the cases where AI changed the work instead of ending it? Do you turn a possibility into a certainty because the scariest version is the most memorable? And if you believe AI is mostly hype, do you collect its dumbest mistakes and skip the work it already does well?
Either way you can believe you are weighing the evidence when you are really picking the evidence that protects the answer you already picked.
F. “Find the facts before you come to any conclusion.”
The original guide told readers to ask who is speaking, how they are trying to influence us, what their words mean, what their interests are, and whether those interests match ours.
Back to the job warning. What does “coming for your job” actually mean? It could mean AI wipes out the whole occupation. It could mean it takes over some of the tasks. It could mean employers need fewer people. It could mean people who use AI get an edge over people who do not.
Those are four different claims. Evidence for one does not prove the others. This is where Petro brings numbers, and this is the part I checked myself against the original report.
Generative AI is the kind of AI that makes new text, pictures, or code instead of just sorting data. A language model is the text kind.
The International Labour Organization published a report on May 20, 2025 called Generative AI and Jobs: A Refined Global Index of Occupational Exposure. One in four workers worldwide has a job that generative AI could partly do. Only 3.3 percent of global employment falls into the highest exposure group. The researchers said AI is more likely to change jobs than to delete them, because most jobs still have tasks only a person can do.
That is not a promise your job is safe. Some people will be displaced. Pay will move. Some fields will need fewer people. What the facts do not support is the simple version, that AI is about to replace nearly everyone. They support a messier version: AI will change different tasks, jobs, and workers in very different ways.
Petro’s point is the discipline, not the number. Do not gather facts about a claim until the claim is sharp enough that facts could prove it wrong.
G. “Guard always, finally, against omnibus words.”
An omnibus word is a big word that carries several meanings at once. The Institute’s own line was that these words “carry all sorts of meanings to the various sorts of men.” Freedom, patriotism, justice, extremism, traditional, woke, and propaganda itself all work this way.
A speaker can get a strong reaction out of the word while never saying what he means by it. The Institute’s fix was sharp definitions. Petro’s version is a sentence you force yourself and the speaker to finish: “By this word I mean…”
The other list of seven
This is the part of the video I did not know. The Institute put out two lists of seven.
The famous one is the seven propaganda devices: name calling, glittering generalities, transfer, testimonial, plain folks, card stacking, and bandwagon. If you went to an American high school, you may have circled one of those on a worksheet. They are easy to teach and easy to test.
Card stacking is leaving out the facts that hurt your side and showing only the ones that help. Glittering generalities are fine-sounding words with no clear meaning attached. Transfer is borrowing the respect people feel for one thing and moving it onto another.
Name calling is sticking a bad label on something so people reject it without looking. Testimonial is borrowing a famous person’s voice. Plain folks is a speaker dressing up as an ordinary person like you. Bandwagon is saying everyone already agrees, so you should too.
The ABCs are the second list, and Petro argues they go deeper. The devices help you examine what someone else is doing to you. The ABCs make you examine what your own identity is doing to the evidence.
Was it really in classrooms?
By the Institute’s own count, its materials were mailed to thousands of high schools, colleges, and public libraries, it corresponded with about 2,500 teachers, and around a million school children were using its methods by the late 1930s. All three of those numbers are the Institute reporting on itself.
One classroom is on the record, and it is a good one. In January 1939, Helen I. Davis, an English teacher at De Witt Clinton High School in the Bronx, wrote up her experiment in The English Journal. Her students read newspaper stories and letters to the editor. They fact-checked a student letter printed in Scholastic called “In Defense of Hitler,” wrote replies, picked the best ones, and sent them to the magazine.
Those students were not memorizing a definition to repeat on a test. They were checking a live message, answering it, and joining the argument around it. One teacher’s write-up is not proof that a million children did the same.
Why it stopped
The video says the guide “quietly faded.” The record reads differently. The Institute ran out of money, got investigated by Congress, and shut itself down in the middle of a war.
The last issue of its bulletin came out on January 9, 1942. The headline was “We Say Au Revoir.” The board had voted to suspend operations, and the editors explained why. They wrote that publishing calm analyses of all kinds of propaganda, “good” and “bad,” “is easily misunderstood during a war emergency,” and that the work “could be misused for undesirable purposes by persons opposing the government’s effort.” They also warned that wartime usually brings a rise in intolerance.
The Dies Committee was a House committee set up to investigate Americans for suspected disloyalty. HUAC is the later and better-known name for that kind of committee.
The Dies Committee investigated the Institute. Its members attacked the Institute’s work from the right. By Miller’s own account, the newspaper publisher William Randolph Hearst told Teachers College that Miller should back off. Teachers College put Miller on leave in 1943 and never brought him back. It fired him in 1948, and told him it was a departmental restructuring. Miller hired a lawyer and sued for $100,000.
You will read in some places that in 1947 a congressional committee formally listed the Institute as a Communist front. I could not confirm that in the scholarly account I used, so I am leaving it out rather than passing it on.
The Institute itself said the cause was money, not the war. I would not lean on that one either. It is the organization’s account of its own failure. Which, given what the guide is about, is worth noticing.
What I am taking from it
I write about AI tools and I use them all day. Here is what the seven letters turn into for me.
Find the fight first. When someone tells me a model is about to replace a whole job category, I want to know who gains if I believe it. That question takes ten seconds and it changes how I read the rest.
Check my reaction before my argument. I have opinions I like. Some of them came from my job, my crowd, and the feeds I read. Letter D does not say those opinions are wrong. It says I should know which ones I tested.
Run it on my side, not just theirs. Letter E is the one I will fail at most often. It is easy for me to spot loaded language in a vendor’s benchmark chart. It is harder to notice when I pick the benchmark that flatters the stack I already run.
Make the claim sharp before I measure it. “AI will take your job” cannot be tested. “This model handles 63 percent of the guesses in my draft layer” can be. That is letter F, and it is also just engineering.
Define the big words or drop them. I use the word “agent” a lot. So does everyone. Half the arguments I read about agents are two people using the word to mean different things. Letter G says finish the sentence.
And turn it on myself. Under the Institute’s own definition, this post is propaganda. I am trying to move what you think, toward an end I picked before I started writing. So was the Institute, which wanted schools to teach its method. Knowing that does not make any of it false. It means the check runs on my side too.
The video closes on this: the schools may have stopped handing out this guide, but nothing stops you from handing it to yourself. That includes the parts that point at you.
Sources
- Stephen Petro, This Curriculum Quietly Vanished From Schools, YouTube, November 16, 2026.
- Institute for Propaganda Analysis, “Some ABCs of Propaganda Analysis,” 1937. Reprinted in Alfred McClung Lee and Elizabeth Briant Lee, The Fine Art of Propaganda, 1939.
- Anya Schiffrin, “Fighting Disinformation in the 1930s: The Institute for Propaganda Analysis and the Reception of Clyde Miller’s Work,” International Journal of Communication 16, 2022, pages 3715 to 3741. Source for the founding, the Filene grant, the Debs story, the farewell issue, the Dies Committee, and Miller’s leave and firing.
- Renee Hobbs and Sandra McGee, “Teaching about Propaganda: An Exploration of the Historical Roots of Media Literacy,” Journal of Media Literacy Education 6(2), 2014, pages 56 to 67. Source for the exact ABC wording, the reach numbers, Helen I. Davis, and the Scholastic letter.
- Helen I. Davis, “Propaganda Enters the English Classroom,” The English Journal 28(1), January 1939.
- Clyde R. Miller, “The Man I Sent to Prison,” The Progressive, October 1963. Miller’s own account of the Debs case.
- International Labour Organization, Generative AI and Jobs: A Refined Global Index of Occupational Exposure, ILO Working Paper 140, May 20, 2025.