Truth in the Feed — Critical Thinking 8 (C1)

CRITICAL THINKING · LESSON 8

Truth in the Feed

C1 (B2–C2) · 90 minutes

Before you scroll on

Task Give each statement a number. Write the four numbers down and keep them.

1

strongly agree

2

mostly agree

3

on the fence

4

mostly disagree

5

strongly disagree

I am better than most people at spotting false information online.

Most of the misleading content I see online is simply false.

If something is true, it is fine to share it.

Algorithms decide most of what I see, and they mostly show me what I already believe.

READING

Not fake, not news

"Fake news" became a global buzzword in 2016 and 2017; Collins Dictionary made it its word of the year in 2017. It is a poor tool for thinking. Politicians of every stripe use it to dismiss reporting they dislike, and much of the problem is neither fake nor news. In a 2017 report for the Council of Europe, Claire Wardle and Hossein Derakhshan proposed a more careful vocabulary for what they called information disorder, built on two questions: is the content false, and is it meant to harm?

Misinformation is false content shared without the intention to harm: an uncle forwarding a bogus health tip because he believes it. Disinformation is false content created or shared knowingly, to deceive or to harm: a fabricated "leaked memo" about a rival company. Malinformation is genuine information used to cause harm, often by dragging something private into public view: a real home address, a real medical record, a real private message.

READING · CONTINUED

Not fake, not news

The categories describe acts, not items. A piece of disinformation becomes misinformation the moment a sincere person passes it on.

Manipulated media sharpens the problem. A deepfake uses machine learning to generate or swap faces and voices. In early 2024 an employee in the Hong Kong office of the engineering firm Arup transferred about HK$200 million (roughly US$25 million) after a video call in which the "chief financial officer" and other colleagues were digital fakes.

Yet Britt Paris and Joan Donovan argued in 2019 that cheap fakes matter at least as much: real footage slowed down, edited selectively or re-captioned with a new time and place. No machine learning required, only a misleading caption.

Critics of the framework ask who decides what counts as "harm", and warn that a label like malinformation can be used to silence true but inconvenient information.

Not fake, not news

False

Intent to harm

Misinformation

Disinformation

Malinformation

Who meant what?

Task Label each case Mis, Dis or Mal. If it depends, say on what.

A

Ana forwards a message to her family group saying that hot water with lemon cures flu. She hopes it helps.

B

A restaurant owner writes fake reviews accusing a rival restaurant of giving customers food poisoning.

C

Someone posts a famous athlete's real, private medical records to embarrass her before a final.

D

A real photo of a 2015 flood is reposted as "this morning's storm".

E

A satirical website publishes a joke headline. Kwame shares it as real news.

If the damage is the same, does the intention behind it matter?

Verification challenge (1)

Task For each post, decide: share, don't share or check further. Write the one clue that decided it. For "check further", write exactly what you would check, and how. About 3 minutes per post.

@NaturalHealthTruths

POST A

@GlobalAlert24

POST B

48,200 followers · 2 hrs ago

1.3M followers · joined March 2026 · 40 min ago

Swedish scientists have PROVEN it!! One teaspoon of turmeric a day cuts your cancer risk by 87%. Doctors won't tell you, because there's no money in it. Share before this gets taken down!

BREAKING: Dam collapses in northern Italy this morning. Thousands fleeing their homes. The media is SILENT. Watch before it's deleted.

wellnesswire-news.co

A 30-second video of water bursting through a concrete dam wall. No place name, no date.

Swedish study: turmeric slashes cancer risk

Top comment, from Tomás R.: "Isn't this the video from 2019?"

4,100 shares

share · don't share · check

share · don't share · check

Verification challenge (1)

@LeilaHaddadOfficiaI

POST C

3,200 followers · joined last week · 1 hr ago

A screenshot of a handwritten note on pink paper.

After ten years, I'm done with music. Thank you for everything. L.

11,000 likes

share · don't share · check

The singer Leila Haddad has 20 million followers on her usual account.

Verification challenge (2)

Task Same as before: share, don't share or check further, with the one clue that decided it. Then circle the post you two found hardest to decide.

Sunrise Coast Health Authority

@RefWatchFC

@SunriseCoastHealth · joined 2014 · yesterday, 16:05

POST D

210,000 followers · 3 hrs ago

POST E

Measles cases have risen in three of our districts this month. Free vaccination clinics will now also open on Saturdays, 9:00 to 14:00, until 30 November. Clinic addresses and opening times are on our website (link). Questions: call our helpline, open weekdays 8:00 to 18:00.

This is the referee who robbed us last night. Here is his home address and his wife's mobile number. You all know what to do.

A photo of a house, with the street name visible.

Comment from Mei L.: "I checked, it's definitely his street."

share · don't share · check

share · don't share · check

1. Which post was quickest to decide? Which clue made it quick?

2. Did "check further" ever really mean "share it later"?

READING

How the professionals check

Sam Wineburg and Sarah McGrew of the Stanford History Education Group asked 45 skilled people to judge unfamiliar websites: 10 historians with PhDs, 10 professional fact-checkers and 25 Stanford undergraduates. In one task they compared pages from two medical organisations with similar names: a large, long-established professional body and a much smaller advocacy group. Many historians judged both credible; many students preferred the smaller group's page. The fact-checkers quickly worked out

who was behind it.

The difference was method. Historians and students read vertically: they stayed on the page, studying the "About" section, the design, the logo, the references. All of these are cheap to fake. Fact-checkers read laterally: after a quick look, they left the site, opened new tabs and asked what the rest of the web said about the source. They also showed click restraint, scanning a page of search results before choosing where to click.

READING · CONTINUED

How the professionals check

They reached sounder conclusions, in a fraction of the time.

The digital-literacy researcher Mike Caulfield turned these habits into four moves he calls SIFT:

1. Stop. Before you read or share, ask whether you know the source. If you get lost while checking, stop again and remember your purpose.

2. Investigate the source. Know who is behind it, with what expertise and agenda, before you engage with it.

3. Find better coverage. When the claim matters more than the source,

look for trusted reporting or expert consensus on it.

4. Trace claims, quotes and media to their original context. Online material is often stripped of it.

One caution. Lateral reading is only as good as what you find when you leave the page. If every source you consult shares one slant, a quick check becomes a quick confirmation.

How the professionals check

Vertical

Lateral

Tab 1

Tab 2

Tab 3

Strategy comparison

Task Copy the table and fill in one row per post (A to E). Then agree on the post where SIFT does not give you the answer, and say why.

Post

What we did: V (stayed on the post) or L (checked elsewhere)

The SIFT move that would settle it fastest

Would we now change our decision?

A

B

C

D

E

1. Which clues in the posts would be easiest for a skilled faker to remove?

2. Fact-checkers read less and decided faster. Is that a lesson for everyone, or only for experts?

READING

Bubbles and chambers

In 2011 the American writer and internet activist Eli Pariser described the filter bubble: a personal universe of information that recommendation algorithms build for each of us, invisibly and without our choosing. His opening example: two friends searched Google for "BP" during the 2010 Deepwater Horizon oil spill; one was shown investment news, the other news of the spill. A related idea, the echo chamber, describes a

space where people mainly hear their own views repeated and rival views are kept out, by algorithms or by their own choices.

Both ideas feel true. The evidence is more mixed. In a 2015 study of 10.1 million US Facebook users, researchers employed by Facebook found that the ranking algorithm did narrow what people saw, but people's own choices, of friends and of what to click, narrowed it more.

READING · CONTINUED

Bubbles and chambers

A 2022 review by the Reuters Institute at Oxford concluded that echo chambers are much less widespread than often assumed: in the UK, for example, an estimated 6 to 8 per cent of the public are in partisan online news echo chambers. It found no support for the filter-bubble hypothesis; platforms generally led to slightly more diverse news use. In 2023 a large experiment during the US presidential election

cut users' exposure to like-minded sources on Facebook by about a third. Their political attitudes did not measurably change.

The research has limits. Much of it comes from the United States, which the Reuters review calls an outlier, and some of it depends on data and cooperation from the platforms themselves.

Talking it over

Task Choose three or four questions. Spend most of your time on the one your group disagrees about.

1. Who should remove false content: platforms, governments, courts or nobody? What happens where you live?

2. Is sharing misinformation by mistake a moral failing, or just bad luck?

3. Should "malinformation" be a category at all, if the information is true?

4. If deepfakes become perfect, will people trust video less, or just believe what suits them?

5. The evidence suggests filter bubbles are rare. Why does the idea remain so popular?

6. Where you live, whom do people trust more: national news, social media or family group chats?

Design your information diet

Task (Individual) Write your plan in your own notes: three rules, plus one source you'll add and one you'll treat with more care. At least one rule must be about sharing, and one about checking.

Task (Pairs) Compare plans. Find one rule of your partner's you would steal, and one you think won't survive a busy week. Say why.

My information diet

Rule 1 (sharing):

Rule 2 (checking):

Rule 3 (my choice):

One source I'll add:

One source I'll treat with more care:

Private reflection

Task Choose one prompt and write privately. Nobody will read what you write. Sharing afterwards is optional.

Prompt 1

Think of a time you, or someone you know, believed or shared something that turned out to be false. What made it convincing?

Prompt 2 (lighter)

Which SIFT move are you most likely to skip when you're busy? What would make you do it anyway?

Back to the start

Task Score the four opening statements again. Has any number changed? Which reading or post moved it?

1

strongly agree

2

mostly agree

3

on the fence

4

mostly disagree

5

strongly disagree

I am better than most people at spotting false information online.

Most of the misleading content I see online is simply false.

If something is true, it is fine to share it.

Algorithms decide most of what I see, and they mostly show me what I already believe.

Before Lesson 9

S

This week, run SIFT on one post you would like to be true.

Note which move settled it, how long it took, and whether you still wanted it to be true afterwards.

I

Next lesson: Is Critical Thinking Universal? Are the tools of this module everyone's, or one culture's habits dressed up as logic?

F

T

TEACHER'S REFERENCE · SKIP OR DELETE BEFORE CLASS

Answer key (1)

WHO MEANT WHAT? suggested

• A. Mis. False, shared sincerely, meant to help.

• B. Dis. Invented, deliberately, to damage a rival.

• C. Mal. True information, made public to cause harm.

• D. It depends. Mis if the person reposting believes it; Dis if they know it is an old photo. Either way it is a cheap fake: a real photo, false context.

• E. Mis. Satire is not disinformation in itself: in Wardle's seven types of mis- and disinformation, satire or parody has no intention to cause harm but has the potential to fool. The harm comes when it travels without its context.

VERIFICATION CHALLENGE suggested; "check further" is acceptable wherever a pair names a specific check

• Post A: don't share. Clues: "PROVEN" and capitals; a suspiciously precise 87%; no named study, university or journal; "doctors won't tell you"; "share before this gets taken down"; an unfamiliar outlet. Investigate the source (what is wellnesswire-news.co?) and Find better coverage (would a real Swedish finding of this size be reported by any health service or major outlet?). Trace claims: research on curcumin, turmeric's main active compound, comes mostly from laboratory and animal studies; human trials on cancer prevention are few, small and mixed, so nothing has been "proven".

• Post B: don't share, or check further. Clues: an account a few months old with 1.3 million followers; no place or date; "the media is SILENT" (a real dam collapse would be on every Italian news site within minutes); the comment about 2019. Trace media: a reverse image search on a still from the video; Find better coverage: Italian news and civil-protection sources. A likely cheap fake: real footage, new caption.

TEACHER'S REFERENCE · SKIP OR DELETE BEFORE CLASS

Answer key (2)

• Post C: don't share. Clues: the handle ends "OfficiaI" with a capital I, not "Official"; 3,200 followers against 20 million on the real account; joined last week; a screenshot, not a link. Investigate the source: check her usual account and any reputable entertainment coverage.

• Post D: share (after a quick check). Clues in its favour: an official account active since 2014; calm tone; specific, checkable details (days, times, end date); a link to the authority's own website and a helpline. Investigate the source: confirm the account is the authority's real one, for example from the authority's website.

• Post E: don't share. This may be entirely true, and SIFT would not show otherwise: it is malinformation. Clues: private details of the referee and of his wife, who is not involved at all; "you all know what to do" invites harassment. Accuracy is not the only test before sharing.

STRATEGY COMPARISON suggested fastest moves

• A: Find better coverage, or Investigate the source.

• B: Trace claims, quotes and media (reverse image search).

• C: Investigate the source.

• D: Investigate the source (is it really the authority?).

• E: SIFT does not settle it: the question is harm, not truth.