L1
·
Quiz
·
Lab
L2
·
Quiz
·
Lab
L3
·
Quiz
·
Lab
L4
·
Quiz
·
Lab
Module Test
Module 4 · Lesson 1

Headlines That Lie Without Lying

How real words get arranged into false impressions
If every word in a headline is technically true, can it still mislead you?

In August 2021, several outlets ran headlines reading "FDA approves COVID vaccine" — technically accurate, since the FDA had approved Pfizer's Comirnaty. But the widely available vaccine at the time was still under Emergency Use Authorization. The two products were legally distinct. Millions of people read one thing and concluded another.

The Misleading Headline Problem

Misleading headlines are one of the most effective tools in information manipulation — precisely because they hide behind technical truth. A sentence can be grammatically and factually correct on its own while still creating a completely wrong impression in the reader's mind.

AI tools can generate, amplify, and detect these patterns at scale. Understanding how headline manipulation works is now an essential media literacy skill.

59%
of links shared on social media are never clicked — people form opinions from headlines alone
(Columbia/French National Institute, 2016)
6x
more likely — false news stories spread faster and wider than true ones on Twitter
(MIT Sloan, Vosoughi et al., 2018)

Five Classic Misleading Headline Patterns

Researchers at the Duke Reporters' Lab and First Draft News have catalogued recurring structures that distort without technically lying:

1. The False Implication A true fact is paired with context that implies a false relationship. Example: "Crime rises in city after mayor endorses police reform" — the sequence is real, the causal link is invented.
2. The Strategic Omission A key qualifier is left out. "Study links coffee to cancer" omits that the study involved mice given 50x normal doses. The study exists; the headline buries what matters.
3. The Outdated Truth A headline accurately describes a past situation but is recirculated as if current. In 2020, images and articles from 2013–2018 border events were widely reshared as if depicting that week's news.
4. The Quote Distortion A real quote is decontextualized. "Senator: We should burn it all down" might be an accurate transcription of a metaphor about bureaucracy — stripped of the surrounding sentences.
5. The Passive-Voice Accusation "Official accused of fraud" requires only that one anonymous person made an accusation. It implies wrongdoing without requiring evidence or rebuttal.

How AI Amplifies the Problem

Large language models trained on news corpora learn these patterns deeply. When asked to "write a compelling headline," they often reproduce the most engagement-optimized versions — which tend to be the most misleading. A 2023 study by NewsGuard found that when given a prompt containing a false premise, GPT-4 produced headlines reinforcing that premise in 80% of test cases without correcting it.

AI-powered content farms — like the network of over 1,000 sites identified by NewsGuard in 2023 — use models to generate headlines at volume, then A/B test them algorithmically to find which phrasings drive the most clicks. The winning headlines are those that produce the strongest emotional reactions, not the most accurate summaries.

Documented Case — 2023

NewsGuard's "Unreliable AI" report (March 2023) found 49 AI-generated news sites spreading false or misleading content. These sites used AI to produce hundreds of articles per day, with no editorial oversight. Headlines were written to imply scandals, health crises, or political wrongdoing that either didn't occur or were wildly exaggerated. Several had strong Google search rankings within weeks of launch.

Reading Strategy: The Headline Test

Before sharing or reacting to any news headline, media researchers at the News Literacy Project recommend three rapid checks:

Check 1 What is it claiming, exactly? Identify the specific assertion — not the emotional impression, but the factual claim being made.
Check 2 What would have to be true for this headline to be misleading? Think about what qualifiers, missing context, or timing issues could make an accurate sentence still deceive.
Check 3 Does the body of the article support the headline? A 2020 Columbia Journalism Review analysis found that in roughly 1 in 5 fact-checked viral articles, the article's own text contradicted or heavily qualified its headline.
Key Insight

The goal of a misleading headline is not to survive a careful reading — it's to travel fast before careful reading happens. Your defense is to pause before sharing, not just before believing.

Module 4 · Lesson 1

Quiz — Headlines That Lie Without Lying

Three questions · Select the best answer
A headline reads: "FDA approves COVID vaccine." The article reveals that the approved product (Comirnaty) is different from the one actually being administered under Emergency Use Authorization. This is an example of which misleading pattern?
Correct. The FDA approval was real, but the omission of the key fact that it applied to a different product formulation than what was being distributed created a false impression about available vaccines.
Not quite. The headline omits a crucial qualifier — that the approved product differed from what was actually being administered. Strategic omission is the pattern here.
According to the MIT Sloan study by Vosoughi et al. (2018), how much faster did false news spread compared to true news on Twitter?
Correct. The 2018 MIT study found false news was six times more likely to be retweeted than true news, and reached people faster and more broadly.
The MIT study found false news spread six times faster. Its novelty and emotional charge made it more shareable than accurate information.
The News Literacy Project's three-check strategy says: before the body of the article, you should ask "What would have to be true for this headline to be misleading?" Why is this check important?
Correct. This check builds the habit of thinking beyond literal truth — asking what qualifiers, missing context, or framing choices could make an accurate sentence still deceive.
This check is specifically about spotting the gap between literal truth and implied meaning — not about bias, outlet reputation, or quote accuracy.
Module 4 · Lab 1

Dissecting Misleading Headlines

Interactive AI Lab · Minimum 3 exchanges to complete

Your Task

In this lab, you'll practice identifying misleading headline patterns with an AI guide. Submit real or invented headlines and the AI will help you classify the manipulation technique, identify what's missing, and suggest how a more accurate headline would read.

Try asking: "Analyze this headline: 'Scientists link red meat to early death'" — or submit any headline you've seen recently and want to examine.
Headline Analysis Lab
L1
Welcome to the Headline Analysis Lab. I'm here to help you dissect misleading headlines and identify the techniques being used. Submit any headline — real or invented — and I'll walk you through what makes it potentially misleading, what information is missing, and how a more accurate version might read. What headline would you like to examine?
Module 4 · Lesson 2

Context Collapse and the Cut-and-Paste Clip

How removing context from video, audio, and quotes changes their meaning entirely
Can a video be completely real and still completely mislead you?

In January 2019, a short video clip circulated showing a teenager in a MAGA hat appearing to stand face-to-face with a Native American elder at the Lincoln Memorial. The clip went globally viral within hours. Hundreds of verified accounts called for the student to be expelled or worse. Full footage released the next day — running over an hour — showed a significantly different sequence of events. By then, the short clip had been seen by tens of millions of people; the correction reached a fraction of that audience.

What Is Context Collapse?

Context collapse occurs when information designed for one audience or situation is stripped of its surrounding context and presented to a different audience as self-explanatory. The term was introduced by media scholar Danah Boyd in the early 2000s, initially to describe social media dynamics. It now applies directly to the spread of misleading news clips.

When a video clip is cut from a longer recording, several layers of context can be lost: what happened before and after, who the audience was, what the speaker's tone meant in full, what questions preceded the statement, and what agreements or disagreements occurred off-screen.

Documented Example

In 2020, a video clip of Nancy Pelosi tearing a copy of President Trump's State of the Union address was edited to remove the full applause context around it and redistributed with a caption framing it as a spontaneous tantrum. Fact-checkers confirmed the act happened but the timing and framing radically altered the perceived meaning. The edited version received roughly 3x the engagement of the full-context version.

The AI Supercharger

AI tools accelerate context collapse in three ways: speed, targeting, and refinement.

Speed: AI-powered clipping tools can automatically identify and extract emotionally charged segments from long videos. A 90-minute press conference can be scanned, the most controversy-generating 30 seconds identified, and a clip prepared for upload within minutes — no human editor required.

Targeting: Recommendation algorithms (which use AI to predict engagement) amplify clips that generate outrage, because outrage produces comments, shares, and watch time. A 2021 Facebook internal memo, later cited in the Wall Street Journal's "Facebook Files" reporting, acknowledged that the algorithm was "increasing the distribution of content with the highest emotional response" — which disproportionately included misleading clips.

Refinement: AI can also be used to generate misleading captions automatically, translating the emotional content of a clip into a caption optimized to drive specific reactions. In 2022, researchers at Stanford Internet Observatory documented several influence operation networks doing exactly this.

The Facebook Files — WSJ, 2021

Internal Facebook research, reported by the Wall Street Journal in September 2021, showed that the company's own engineers had identified that its content-ranking algorithm pushed users toward increasingly extreme and emotionally charged content. Out-of-context clips were consistently among the highest-performing content types. Internal proposals to reduce their spread were repeatedly deprioritized.

The Covington Case: A Timeline

The January 2019 Covington Catholic High School incident became a landmark case study in how context collapse operates in real time:

Day 1 — Within Hours
A 2-minute clip circulates on Twitter. Framing: a student is blocking an elder. High-profile accounts share with condemnations. The clip receives millions of views.
Day 2
Full 1-hour-46-minute video surfaces. It shows the encounter was preceded by aggressive shouting from a third group and the student held his ground rather than advancing. Context changes the story significantly.
Days 3–7
Major outlets issue corrections and extended analyses. Several journalists delete original tweets. The student's family files defamation suits against multiple outlets. Washington Post settles for a reported $250 million.
Study — 2020
A Shorenstein Center analysis found the corrective coverage reached approximately 15% of the audience that saw the original clip — a ratio consistent with prior research on correction uptake.

Detection Skills: Watching for Cuts

When evaluating any video clip shared as news, professional fact-checkers at organizations like PolitiFact and AFP Fact Check use the following signals to detect potential context collapse:

Abrupt starts/ends Clips that begin mid-sentence or end before a reaction often signal that surrounding content would change the meaning.
Mismatch between caption and content When the caption describes a strong action ("attacks," "slams," "destroys") but the video shows a more ambiguous exchange, the caption is doing interpretive work the video doesn't support.
No original source link Clips without a link to the full recording, or where the full recording has been removed, should be treated with high suspicion.
Emotional reaction in comments before broad viewing If thousands of people are already furious about a clip that posted 30 minutes ago, the emotional framing may be doing more work than the footage itself.
Key Insight

A real video is not the same as a true account. Context is not decoration — it is meaning. A clip that accurately captures 30 seconds while omitting 90 minutes is not an honest representation of an event.

Module 4 · Lesson 2

Quiz — Context Collapse

Three questions · Select the best answer
The term "context collapse" was introduced by media scholar Danah Boyd to describe what phenomenon?
Correct. Boyd's concept describes how content loses its original contextual meaning when transplanted to a new audience — a dynamic that now drives much viral misinformation.
Context collapse refers to stripping information of its original context when presenting it to a new audience. The other options describe different (also real) problems.
In the Covington Catholic High School incident (2019), a Shorenstein Center analysis found that corrective coverage reached approximately what percentage of those who saw the original viral clip?
Correct. Only about 15% of the original clip's audience saw the corrective coverage — a ratio consistent with other research showing corrections almost always reach far fewer people than the original false story.
The Shorenstein Center analysis found only about 15% of the original audience saw the correction. Corrections routinely reach only a fraction of the original audience.
According to the Facebook Files reporting by the Wall Street Journal (2021), what did Facebook's own internal research show about its content-ranking algorithm?
Correct. The Facebook Files showed internal engineers had identified the problem but business priorities repeatedly delayed fixes — a structural case study in how platform incentives can amplify misleading content.
The Facebook Files showed the opposite — the algorithm boosted emotionally charged content, and internal proposals to address it were deprioritized for business reasons.
Module 4 · Lab 2

Context Collapse Detective

Interactive AI Lab · Minimum 3 exchanges to complete

Your Task

Practice identifying context collapse in video clips and quotes. Describe a clip or situation you've encountered (or one from this lesson), and the AI will help you identify what context is missing, what signals suggest manipulation, and how to find the full original source.

Try asking: "A clip is going viral showing a politician saying 'we need to cut every program' — it's 8 seconds long. What context clues should I look for?" Or describe any real clip you've seen.
Context Collapse Lab
L2
Welcome to the Context Collapse Detective Lab. I'll help you develop skills to identify when a video clip or quote has been stripped of essential context. Describe a clip, a quote, or a scenario — real or hypothetical — and I'll walk you through the warning signs of context collapse and how to locate the original full-length source.
Module 4 · Lesson 3

Fabricated Quotes and Synthetic Sources

When AI invents the evidence — and how to catch it
What happens when the "expert" who said something never existed?

In June 2023, a federal judge in New York sanctioned attorneys who had submitted a legal brief citing cases that did not exist. The cases had been generated by ChatGPT and presented as real precedents. One fake citation was "Varghese v. China Southern Airlines Co." — the case never happened. When the opposing counsel couldn't find the cases and the judge demanded copies, the attorneys submitted more AI-generated text purporting to be the case opinions.

The Fabricated Quote Problem

Large language models generate plausible-sounding text — including plausible-sounding quotes from real people. Because these models have ingested vast amounts of text where real people expressed views, they can generate statements that sound like how a specific person would speak, using vocabulary and sentence structures consistent with their documented writing. But the specific quote was never said.

In 2023, researchers at the University of Oregon tested GPT-4 by asking it to produce quotes from named journalists and academics. In over 60% of cases, GPT-4 produced convincing, contextually appropriate quotes that were entirely fabricated. None were labeled as such.

~60%
of AI-generated "quotes" from named public figures were fabricated but plausible in University of Oregon testing, 2023
8
fake legal citations submitted in the Mata v. Avianca case (2023) — none were real court decisions

Fabricated Experts: The Synthetic Source

Beyond fabricating quotes from real people, AI can generate entirely fictional experts: fake professors, fake studies, fake institutions. These synthetic sources are especially dangerous because standard verification methods — searching for the quote — fail when the person doesn't exist.

NewsGuard — Documented Case, 2023

NewsGuard's 2023 analysis of AI-generated news sites found multiple articles citing "Dr. James Harmon, infectious disease specialist at Johns Hopkins" — a person who does not appear in any Hopkins faculty directory or academic database. The same name appeared in articles from at least four different AI-generated sites, suggesting the model had learned to generate plausible-sounding medical authority figures using real institutional names.

The Mata v. Avianca Case — A Deep Look

The 2023 New York sanctions case became the most widely reported example of AI hallucination producing real-world harm in a legal context. The sequence matters:

Step 1
Attorney used ChatGPT to research case precedents for a personal injury suit against an airline. ChatGPT produced six cases with realistic-sounding citations, docket numbers, and dates. None existed.
Step 2
When asked to confirm the cases were real, ChatGPT said yes. When asked for the full opinions, it generated plausible-sounding opinion text. The attorney accepted this output without independent verification.
Step 3
The brief was filed in federal court. Opposing counsel couldn't find the cited cases. The judge ordered copies. The attorney submitted AI-generated summaries as if they were real opinions.
Outcome
Judge P. Kevin Castel sanctioned the attorneys $5,000 each. The case was dismissed. The incident was covered globally and led to multiple bar associations issuing AI-use guidelines for legal practice.

Why AI Fabricates Confidently

Language models don't know what they don't know. They generate the most statistically likely continuation of text — and for a question like "What did expert X say about Y?", the most likely continuation is a plausible-sounding quote in that expert's style, because the training data contains many such quotes. The model has no internal flag for "this is a fabrication." It generates with the same confidence whether retrieving something accurate or inventing something entirely new.

This is not a bug being actively fixed — it is a fundamental property of how current language models work. Even models with retrieval augmentation (the ability to search databases) can generate confident but fabricated text when the retrieval step fails silently.

Key Warning Sign

If an article quotes an expert you've never heard of, from an institution you recognize, making a very precise and conveniently relevant claim — that specific combination is a warning sign. Check the expert's name in the institution's official faculty directory, in Google Scholar, and in general web search. If they don't appear in at least two independent sources, treat the quote with high skepticism.

Verification Protocol for Quotes

Professional fact-checkers use a layered protocol when verifying quotes in AI-influenced media environments:

Step 1: Source Check Search the speaker's name across independent sources (university directory, LinkedIn, academic databases, Google Scholar). A real expert will have a traceable presence.
Step 2: Quote Origin Search the exact phrase from the quote in quotation marks. Real quotes that were genuinely said appear in multiple sources. An AI-generated quote typically appears only in the article (or in copies of it).
Step 3: Publication Check Verify the publication or study cited actually exists. For academic papers, check Google Scholar, PubMed, or the relevant journal's website. For legal cases, check PACER or Westlaw.
Step 4: Context Check Even if the source exists, verify the quote appears in the cited document. AI can accurately attribute sources but fabricate the specific words attributed to them.
Key Insight

The most dangerous AI fabrications are not obviously wrong — they are precisely, plausibly, convincingly right-sounding. The goal is not to make you disbelieve everything; it is to make verification a habit, especially when a quote seems to perfectly support a claim you already believe.

Module 4 · Lesson 3

Quiz — Fabricated Quotes and Synthetic Sources

Three questions · Select the best answer
In the Mata v. Avianca case (2023), what happened when the attorney asked ChatGPT to confirm that the cited cases were real?
Correct. ChatGPT confirmed the fake cases were real and generated more fabricated content when pressed — illustrating that AI models have no internal flag distinguishing accurate retrieval from confident fabrication.
ChatGPT confirmed the fake cases were real and generated fabricated opinion text — it did not warn, refuse, or provide real links. This is why independent verification is essential.
Why is it especially dangerous when AI generates quotes from fictional experts using real institutional names (like a fake professor at a real university)?
Correct. The familiar institution name bypasses skepticism, and the fictional person can't be found through normal quote-tracing methods — making the fabrication harder to catch than a simply wrong fact.
The core danger is that a real institution name lends credibility while the fictional person evades standard verification methods. The quote can't be traced because the source never existed.
According to the lesson, what is the most reliable first step to verify that a quoted expert is real?
Correct. A real expert leaves traces in multiple independent sources. If a name only appears in the article in question and its copies, that's a strong signal of fabrication.
The most reliable check is searching official institutional directories and academic databases — not social media (easily faked), another AI (which might also fabricate), or engagement metrics (irrelevant to truth).
Module 4 · Lab 3

Quote Verification Trainer

Interactive AI Lab · Minimum 3 exchanges to complete

Your Task

Practice verifying quotes and sources. Give the AI a quote, an expert name, or a claim from an article you want to fact-check, and it will walk you through the step-by-step verification process — helping you determine whether the source is real and whether the quote can be traced.

Try: "An article says 'Dr. Sarah Kim, neuroscientist at MIT, warns that social media causes permanent memory loss in teenagers.' How do I verify this?" Or bring your own quote to check.
Quote Verification Lab
L3
Welcome to the Quote Verification Trainer. I'll help you practice the step-by-step protocol for verifying whether an expert is real and whether a quote can be traced to an original source. Share any quote, expert claim, or source you want to investigate — real or hypothetical — and I'll walk you through exactly how to check it.
Module 4 · Lesson 4

When Real News Gets Twisted by Framing

The same facts, two completely different stories
How can two news reports cover an identical event and lead readers to opposite conclusions?

During the summer of 2020, media monitoring organization AllSides tracked coverage of the same protest events on the same days. Fox News and MSNBC ran headlines about identical incidents: Fox's framing emphasized property damage and disorder; MSNBC's framing emphasized crowd size and police response. Neither set of facts was false. The framing made the same event appear to be two different stories — and two different countries.

What Is Framing?

Framing, in communication research, refers to the emphasis, selection, and presentation choices that shape how an audience interprets a story. Introduced by sociologist Erving Goffman in the 1970s and extensively developed by political scientists including Robert Entman, framing theory holds that which facts are selected, which are emphasized, and which are left out all actively construct the audience's understanding — even when no individual claim is false.

Framing is not inherently manipulative — all storytelling involves selection. It becomes a tool of misinformation when it is systematically used to produce misleading interpretations of events, especially when audiences are unaware of what has been left out.

Five Framing Mechanisms

1. Definitional Framing What something is called changes how it's perceived. "Undocumented immigrants" vs. "illegal aliens" both refer to the same legal status but produce measurably different emotional and policy responses — documented by political scientist Valentino et al. in peer-reviewed research.
2. Causal Framing Assigning blame changes the story. A job loss can be framed as a company decision, a government failure, an automation trend, or a worker's inadequacy — the same event, four different implied causes and solutions.
3. Moral Framing Placing events in a moral context (heroism vs. recklessness, justice vs. lawlessness) activates different values in readers. Coverage of the same protest can frame participants as civil rights actors or as threats to public order.
4. Numerical Framing Statistics can frame the same fact very differently. "1 in 100 patients experienced side effects" and "The drug causes side effects in 1% of patients" are identical — but "1 in 100" consistently produces higher perceived risk in study participants (Slovic, 2000).
5. Omission Framing What's left out constructs meaning as much as what's included. Covering a government jobs report without mentioning the methodology change that inflated the numbers is technically truthful reporting that produces a false impression.
Documented Case — Israel-Gaza Coverage, Pew Research 2023

A Pew Research Center analysis of news coverage in 2023 found that the same confirmed casualty figures were framed radically differently across outlets: some led with numbers, others led with who was responsible, others led with comparative historical context. Readers of different outlets, given identical underlying facts, came to substantially different conclusions about proportionality and cause — entirely as a function of framing choices.

How AI Scales Framing Manipulation

AI makes framing manipulation dramatically more efficient and targetable. A human propagandist writing slanted coverage can produce perhaps 5–10 articles per day. An AI system — as documented in Stanford Internet Observatory research on state-sponsored influence operations — can produce thousands of uniquely framed versions of the same story, each tailored to a different ideological audience, within hours of an event occurring.

In 2022, researchers at the University of Washington and Stanford published a study showing that GPT-3 could reliably reframe a neutral news summary into any of several ideologically distinct versions when prompted with simple framing instructions. The output was rated by human evaluators as indistinguishable from professionally written partisan news coverage.

Stanford Internet Observatory — Influence Operations, 2023

A 2023 SIO report documented a Chinese state-linked influence operation using AI to generate English-language content about U.S. domestic issues. The operation produced articles on topics including gun control, racial politics, and immigration that were framed to amplify divisions — all using verifiable facts, just selectively emphasized. The articles were placed on legitimate-looking news sites and shared through authentic-seeming social accounts.

Reading Against the Frame

Media researchers recommend the following practices to identify and counteract framing effects when consuming news:

Cross-source comparison Read coverage of the same event from at least two sources with different editorial orientations. Pay attention not to who disagrees on facts, but who emphasizes different facts.
Ask "what's missing?" After reading any news story, actively ask: What affected group wasn't interviewed? What alternative explanation wasn't mentioned? What time context was omitted?
Label-spotting Notice emotionally charged labels and ask whether a neutral alternative would change your interpretation. This is especially important for terms describing people, groups, and events.
The "so what?" test Ask: given these facts, what is the reader being led to conclude? Is that conclusion supported by the facts, or by the framing around the facts?
Key Insight

No news story is frameless — selection is inherent to storytelling. Your goal is not to find "unframed" news, which doesn't exist, but to recognize the frame in use, identify what it emphasizes and omits, and seek out the perspectives left out of the frame you're reading.

Module 4 Summary

Across this module, you've examined four distinct ways real news gets twisted: misleading headlines that use true words to create false impressions; context collapse that strips video and quotes of the surrounding information that gives them meaning; fabricated quotes and synthetic sources that AI generates with dangerous plausibility; and framing that shapes how audiences interpret identical facts.

In each case, AI either creates the problem, amplifies it, or both. And in each case, the defense is not blanket skepticism but trained, systematic evaluation: checking what's omitted, verifying sources, seeking context, and reading against the frame.

Module 4 · Lesson 4

Quiz — Framing Effects

Three questions · Select the best answer
According to media researcher Robert Entman's framing theory, how does framing produce misleading interpretations even when no individual claim is false?
Correct. Entman's framing theory holds that selection, emphasis, and omission actively construct meaning — so a story can be built entirely from true facts while still producing a fundamentally misleading interpretation through what it chooses to include and exclude.
Framing theory focuses on selection, emphasis, and omission — not images, repetition, or embedded falsehoods. These are other real techniques but not what framing theory describes.
A 2022 University of Washington/Stanford study found that GPT-3 could reliably reframe a neutral news summary into ideologically distinct versions. How did human evaluators rate this AI-generated partisan content?
Correct. Human evaluators could not distinguish the AI-generated framed content from professional partisan coverage — a significant finding about the scale at which framing manipulation can now be deployed.
The study found human evaluators rated the AI output as indistinguishable from professional partisan coverage — making AI-generated framing manipulation nearly impossible to detect by reading alone.
The lesson states that "no news story is frameless." What does this mean for how you should approach news consumption?
Correct. The goal is not to find frameless news (it doesn't exist) but to develop the skill of recognizing frames, identifying their omissions, and actively seeking out the perspectives they leave out.
The lesson specifically rejects blanket distrust and points out numbers can also be framed (the 1-in-100 vs. 1% example). The practical skill is recognizing and reading against frames, not abandoning news consumption.
Module 4 · Lab 4

Framing Analysis Workshop

Interactive AI Lab · Minimum 3 exchanges to complete

Your Task

Practice identifying framing in real news coverage. Share a news excerpt, a story summary, or even just a topic, and the AI will help you identify which framing mechanisms are at work, what perspectives are being emphasized or omitted, and how the same facts might be framed differently.

Try: "Analyze the framing in this headline: 'Protesters Clash With Police Outside City Hall'" — or share any recent news story you want to examine for framing effects.
Framing Analysis Lab
L4
Welcome to the Framing Analysis Workshop. I'll help you identify and analyze framing mechanisms in news coverage — definitional choices, causal attributions, moral framing, numerical framing, and strategic omissions. Share any news excerpt, headline, or story and I'll walk you through what framing choices are being made and what perspectives they include or leave out.
Module 4

Module Test — When the News Gets Twisted

15 questions · 80% required to pass
1. A headline reads: "Scientists Discover Link Between Processed Food and Cancer." The study actually found a weak correlation in a sample of 200 mice given doses 100x normal intake. Which misleading pattern is primarily at work?
Correct. Omitting the animal model, tiny sample, and extreme dosage makes a technically true headline radically misleading about human health risk.
This is strategic omission — the headline omits the crucial qualifiers (animal study, extreme dose, small sample) that would prevent readers from falsely concluding the finding applies to humans.
2. What percentage of links shared on social media are never clicked, according to the Columbia/French National Institute study mentioned in the module?
Correct. Nearly 6 in 10 links are shared without the article being read — which means headlines alone shape the majority of opinion formed from shared links.
The study found 59% of links are never clicked — meaning headlines alone are doing the persuasion work for the majority of shared content.
3. The Covington Catholic High School viral clip case (2019) illustrates which core principle about video evidence?
Correct. The Covington case is a landmark example: the short clip was authentic, but its circulation without the full context created a fundamentally false impression of the event.
The Covington case teaches that authentic video can still mislead when stripped of context — not about deepfakes, media fabrication, or platform regulations.
4. In the context of context collapse, what does an "abrupt start" to a video clip typically signal?
Correct. Clips that start mid-sentence or mid-action are a key warning sign that a longer recording existed and that something important came before.
Abrupt starts suggest context has been removed — the most important signal is that something happened before the clip starts that you are not seeing.
5. The Facebook Files (WSJ, 2021) revealed that Facebook's internal engineers had identified a problem with the recommendation algorithm. What was that problem?
Correct. The Facebook Files showed a structural conflict between engagement-optimization and accuracy — and that business priorities repeatedly delayed proposed fixes.
The Facebook Files documented that the algorithm rewarded emotional content (which skewed toward misleading material) and that engineers' proposals to address it were deprioritized for business reasons.
6. A federal judge sanctioned attorneys in the Mata v. Avianca case (2023). What was the fundamental reason for the sanction?
Correct. The attorneys cited six fabricated cases and then, when caught, submitted more AI-generated fabrications — compounding the initial error. The judge sanctioned them $5,000 each.
The core violation was submitting fictional case citations generated by ChatGPT, which the attorney accepted as real without independent verification, then doubling down with more fabrications when ordered to provide copies.
7. Why is it especially difficult to detect AI-generated fabricated quotes from real people?
Correct. AI trained on a person's documented speech can produce convincing fabrications in their style — which is why searching for the specific phrase in multiple independent sources is essential.
The danger is stylistic plausibility — AI generates text consistent with how the real person typically speaks, making the fabrication hard to detect without independently tracing the quote to an original source.
8. The University of Oregon testing (2023) of GPT-4's quote generation found what result?
Correct. More than 6 in 10 generated quotes were fabricated but plausible — and none were labeled as such. This is why independent verification is essential.
The study found GPT-4 fabricated convincing quotes in over 60% of cases, with no labeling. The model does not flag its fabrications.
9. Robert Entman's framing theory was developed in which academic field?
Correct. Entman developed framing theory within political science, specifically examining how selection and emphasis in news coverage shape audiences' political interpretations of events.
Entman's framing theory was developed in political science, building on Goffman's earlier sociological work on frames.
10. The 2022 Stanford Internet Observatory documented a Chinese state-linked influence operation doing what with AI?
Correct. The operation used true facts — just strategically framed — on sites designed to look like legitimate news outlets, making the content harder to dismiss as obvious disinformation.
The SIO documented framing-based manipulation using verifiable facts — not deepfakes, hacking, or translation of foreign disinformation.
11. Numerical framing research by Slovic (2000) showed that "1 in 100 patients experienced side effects" produced what response compared to "1% of patients"?
Correct. "1 in 100" personalizes the statistic — readers imagine themselves as that one person — producing measurably higher perceived risk than the equivalent percentage.
Slovic's research found "1 in 100" produces higher perceived risk. The fraction format is more concrete and personal, making the risk feel more vivid and significant.
12. NewsGuard's 2023 "Unreliable AI" report identified over how many AI-generated news sites spreading false or misleading content?
Correct. NewsGuard identified a network of over 1,000 AI-generated sites — many with strong search rankings — producing misleading content at scale with no editorial oversight.
NewsGuard identified over 1,000 such sites, demonstrating the scale at which AI can automate misleading content production.
13. When checking a news quote's authenticity, what does it mean if the exact phrase only appears in the article you're reading and direct copies of it?
Correct. Genuine quotes that were actually said accumulate citations over time in multiple independent sources. A quote that appears only in one article and its copies has not been independently corroborated.
Real quotes get cited, discussed, and referenced across multiple independent sources. A quote appearing only in one place and its copies lacks independent corroboration — a key warning sign.
14. The "passive-voice accusation" headline pattern (e.g., "Official Accused of Fraud") is misleading primarily because:
Correct. The passive-voice accusation only requires someone — even one anonymous person — to have made an allegation. The headline implies wrongdoing while the low evidentiary bar protects the outlet legally.
The problem is the near-zero evidentiary threshold: any accusation by anyone produces a technically defensible headline that implies guilt. Evidence and rebuttal are not required.
15. Which of the following best describes the correct approach to news consumption in a world where AI generates framed content at scale?
Correct. The module's consistent conclusion is that the defense against AI-amplified news manipulation is trained, systematic evaluation — not blanket avoidance or blind trust in any single verification tool.
The module explicitly rejects blanket distrust and over-reliance on single tools. The goal is systematic evaluation skills: omission-checking, source verification, context-seeking, and cross-source comparison.