Special investigation · Everyday belief · The information loop
When did you start
believing that?
Something you saw. Something you heard. Something that began to feel obvious. How did it become something you knew, and what did you do next?
You read a post saying a kettle broke after two weeks. Later, a colleague repeats the warning. You own the same model. Is that enough reason to think yours will break too?
8:12 am · The post“My new kettle stopped working after a fortnight.”
Lunchtime · The conversation“Apparently those kettles are unreliable.”
Evening · Your kitchenYou own the same model. Do you now expect it to fail?
The starting claim
At what point did a complaint become a fact?
You read a post: ‘My new kettle stopped working after a fortnight.’ A friend shares it with a warning about the brand. Later, a colleague repeats the warning. You own the same model, and it still works.
Would you now worry that yours is likely to break? Before deciding, ask whether anyone has added new evidence. Did your friend have the same problem? Did your colleague own a broken one?
Trace the accounts back and they all lead to the first post. Three appearances. One complaint. The complaint might be accurate. A fault across the whole model is a separate claim that needs more evidence.
At what point did ‘someone’s kettle broke’ become ‘these kettles are unreliable’? The words changed as the complaint travelled. A conclusion about a whole brand appeared along the way.
Think of a claim you have repeated recently. What did the source actually say, and what did you conclude from it? Keep that claim in mind as you read.
Trace the three appearances
The short answer
Three appearances can still be one witness.
Repeated appearances can make a claim feel more credible without adding independent evidence. Separate what you encountered, what you inferred and what the evidence supports before deciding what to do with it. Your next click or share may also help shape what reaches you or someone else.
Listen, explore or change the Lens
Choose how to read this
Read, listen or follow the question.
The facts do not change. Each view uses the same published sources and leaves the same questions open.
Different viewsSee what each perspective notices—and may miss.
See through another Lens
Which view do you want to understand first?
Each view notices something useful. None is allowed to stand in for the complete evidence.
Everyday reader
A familiar claim still deserves a source.
What this view explainsHow to separate another appearance from another check.
What it may missThe original complaint could still be accurate.
Distribution & selection
Who chose what reached you?
A new heading and photograph can make the same kettle complaint look like a fresh report. Before counting another witness, follow the link back. What is new here: the evidence, or its packaging?
A repeated opinion from one source can affect estimates of how widely it is held; other experiments show an initial online rating can influence later ratings. Neither finding turns popularity into proof. 1213
People, editors and recommendation systems choose what becomes prominent. Australia’s proposed feed choice is one attempt to give users more say in that selection. A screen full of similar posts is not a count of everyone’s opinion. AU_FEED_01
We select too. A review found that people often choose information that fits existing views, although circumstances can also draw them towards challenging material. Would you inspect the source as closely if the claim agreed with you? 14
Six more ways to examine what you believe
From an AI answer to a familiar map: each story offers a different question to try on information you encounter. Read them in any order.
- Artificial IntelligenceDid you ask AI the question, or give it the answer?
- Media & RegulationWho chose what you saw?
- Geography & Visual ModelsNobody told you Greenland was that big.
- Synthetic Media & LawYou saw it. Except it never happened.
- Politics & AccessWhat happens when you change the room?
- Digital Privacy & TrackingYour phone may not need to listen.
Familiarity as a clue
Does it feel true, or have you checked it?
Why might three appearances feel persuasive? Across many experiments, people tend to rate repeated statements as truer than new ones. A 2026 analysis combined 182 studies and found a small average effect after adjusting for small study effects, with substantial variation. It is a tendency in measured tasks, not a rule that you believe every repeated claim. 1
Familiar material can be easier to process, and that ease may serve as one clue to truth. The task and context matter. Repeating a true fact can help you learn it too; familiarity alone settles neither truth nor falsity. 5
Knowledge is useful but does not always prevent repetition effects in experiments. Later work comparing different models cautions against turning that into ‘knowledge is useless’. 342
‘I recognise it’ and ‘I checked it’ are different sentences. Which one describes the claim you brought to this story?
Equivalent descriptions
Same ten parcels. What did the wording invite you to conclude?
Imagine ten deliveries: nine arrive on time and one arrives late. ‘Nine out of ten on time’ and ‘one out of ten late’ describe the same parcels. One sentence points at the success; the other points at the failure.
Framing experiments show that equivalent descriptions can change preferences. These parcels illustrate emphasis; they do not predict how you would choose in a high stakes emergency. 9
Both sentences can be accurate. Neither tells you how late the parcel was, what it carried or how other services compare. What did the wording make prominent, and what would you still need to know before choosing?
An initial number can also influence later estimates, a family of findings called anchoring. ‘Was $180, now $99’ does not by itself show that $99 is good value. 10
Source monitoring
Can you remember the source as well as the story?
A week later, somebody asks about the kettle brand. It would be easy to answer: ‘I read somewhere they’re unreliable.’ But where? Could you take that person back to the evidence?
Research distinguishes familiarity with a statement from recollection of its source. Remembering several examples can also affect judgments about frequency; it cannot tell us the fault rate without knowing how many kettles were sold and how complaints were counted. 1811
Can you find the original source? If you cannot, say ‘I remember hearing this’ rather than presenting it as an established fact.
Misinformation persistence
What would make you reconsider?
Suppose the original poster adds an update: the kettle did not fail. The power point was faulty. Would that update reach everyone who saw the warning? If you had passed it on, would you pass the correction on too?
Research finds that corrected accounts can still affect later reasoning. Corrections can help, especially when they explain what really happened. Here, the faulty power point provides a replacement explanation rather than just an instruction to forget the complaint. 17
Some studies found that people drawn into a story gave answers that fitted it. Ask the same question of this article: does the explanation make sense because the evidence supports it, or because the story is easy to follow? 16
The feedback loop
What will you do with what you now believe?
You could replace the kettle, warn a friend, check independent fault reports or carry on making dinner. Which action would the evidence justify? A report about one appliance and reliable fault data about a whole model support different decisions.
Before sharing, consider what your friend will receive. Your confident warning might sound like another independent check. Could you instead link the original complaint and explain what remains uncertain?
A search, click or follow can also help a service choose what to show next, depending on its design and settings. A click might mean curiosity or scepticism. The system’s prediction need not match your reason for clicking. TRACK_01
An intention is not an action: a review of experiments found that changes in intentions did not lead to equal changes in behaviour. So make the next step concrete. Find the source, look for an independent check, or correct something you have already shared. 31
What you can do
Before you share, ask one question.
Is this another source, or another appearance?
Follow a consequential claim back to its origin. If the later accounts all point to the same post, count them as a trail of distribution, not separate confirmation. If they add independent observations, ask what those observations actually establish.
Research on accuracy prompts and explaining manipulation techniques gives reason to explore such pauses. It does not establish that this article changes behaviour or makes anyone immune to error. 3233
You do not have to distrust everything. You can give a claim as much confidence as its evidence earns, and change that confidence when the evidence changes.
A quick reading activity
Read the activity explanation
About the optional activity
We chose a yellow umbrella.
The activity adds a few illustrations and small details to the story. Those extras are left out when you choose “Keep reading”. This is a reading activity, not a controlled comparison.
You may have thought of the umbrella, another object in the story, or something unrelated.
We cannot tell you what caused your answer. A match could involve the cues, conscious recognition, a recent memory or coincidence. If nothing came to mind, that does not mean you are unaffected. Without a comparison, this activity cannot establish an effect.
Research on semantic priming shows that related ideas can affect responses in controlled word tasks. It does not validate this exact article activity. Some broader claims about unconscious behavioural priming also have contested replication histories. 78
We can show you what we placed. We cannot tell you what your mind did with it.
See the planned cues for the optional version
These describe the optional version; they are not a record of exposure when you chose to read without it. Intended association, effect untested.
- At what point did a complaint become a fact?
A morning kitchen scene with a kettle, mug, and a bright yellow folded umbrella resting on the bench.
Visible object and colour association. Intended association, effect untested.
- Who chose what reached you?
Reuse the bright yellow tag colour as an unobtrusive divider accent.
Visual familiarity; no universal colour psychology claim. Intended association, effect untested.
- Same ten parcels. What did the wording invite you to conclude?
The parcel is waiting under cover, out of the rain.
A related everyday context, separate from delivery statistics. Intended association, effect untested.
- Can you remember the source as well as the story?
A hallway scene with a raincoat and a bright yellow umbrella resting by the doorway.
Repeated object and colour association, openly available to any reader who notices. Intended association, effect untested.
- What would make you reconsider?
A final small yellow accent, followed by no further cue decorations before the pause.
Fixed repetition, not a personalised attempt to force a response. Intended association, effect untested.
We did not ask for or save your answer. This activity cannot tell us what caused it.
Take one belief back to its source.
Pick one claim you have repeated recently. Where did it start? What evidence supports it? What would make you reconsider it?
Then choose one next step: find the original record, check an independent account, or update someone you passed it to. If you cannot establish it, say what you know and what you are still checking.
And what about the things you tell yourself?
There is a line people repeat about their own thoughts: ‘Whatever you say about yourself, your subconscious accepts as fact.’ It sounds like a complete explanation. The evidence below does not establish that universal rule.
That does not make the words you say to yourself irrelevant. A review of sport studies found benefits from exercises that guide how people speak to themselves. Experiments using a person’s own name instead of ‘I’ found differences in responses to social stress. These findings concern performance or regulation in specific tasks, not thoughts becoming facts. 2524
Positive wording is not guaranteed to help either. In brief experiments, some participants who reported lower self esteem felt worse after repeating positive statements. Other research asks people to reflect on important personal values. Reviews found small average benefits from those activities. Repeating praise and reflecting on values are different exercises. 232627
Instead of ‘I’m hopeless at this’, you might ask: ‘I don’t understand this part yet. What can I check next?’ That is a practical check, not medical advice. An unwanted thought does not cause bad luck or illness, and this research does not blame anyone for struggling to think positively.
Why did the original subconscious saying feel familiar to you, if it did? We cannot reconstruct that history from recognition alone. Treat your own explanation for a thought as something to check, too.
Inspect the basis
Sources and limits
Why Lens says this
The finding
Repeated appearances can make a claim feel more credible without adding independent evidence. Separate what you encountered, what you inferred and what the evidence supports before deciding what to do with it. Your next click or share may also help shape what reaches you or someone else.
Direct records
A feed choice was proposed in Australia. Repetition studies measure truth ratings. Advertising guidance describes use of past activity. Behaviour studies keep intentions and actions separate.
Lens interpretation
A claim can reach a person through selection, then reappear. The route and the response must be checked separately.
Evidence that limits a simple verdict
A repeated claim may be true. A chosen feed is not necessarily accurate. A click can reflect curiosity or doubt.
What we checked
Supplied targeted psychology review of 41 source records, one follow-up paper, and the add-on issue source register. This is not an exhaustive systematic search. Source access ranges from catalogue and abstracts to selected full texts. No Lens reader experiment or raw-data reanalysis.
What would change the answer
- A correction or retraction changes a cited finding
- New research materially changes the synthesis
Research cutoff: 30 September 2026.
1. Ye, Attali, Ghazi, Cachia, Cassotti & Borst · 2026
Systematic review and meta-analysis of the evidence for an illusory truth effect and its determinants
meta-analysis
What it supports: Repetition increased perceived truth on average; effects varied considerably.
What it cannot establish: An individual probability of believing a claim; a universal dose-response rule.
Supplied research review: Publisher full-text HTML; abstract, selection, bias and limitations inspected
3. Fazio, Brashier, Payne & Marsh · 2015
Knowledge does not protect against illusory truth
experiments
What it supports: Relevant prior knowledge did not eliminate repetition effects in these tasks.
What it cannot establish: Knowledge is useless, or every known falsehood will be believed.
Supplied research review: Indexed abstract and bibliographic record
5. Hassan & Barber · 2021
The effects of repetition frequency on the illusory truth effect
experiments
What it supports: Tests repetition frequency; discusses processing ease as an explanatory account.
What it cannot establish: Fluency is the only mechanism, or any pleasant layout makes claims true.
Supplied research review: Publisher abstract and research text surfaced
7. Hutchison and colleagues · 2013
The semantic priming project
large experimental dataset
What it supports: Related-word context can facilitate responses in timed word tasks.
What it cannot establish: A colour or a few hints will make readers freely choose one exact object.
Supplied research review: Publisher HTML abstract and design inspected
8. Doyen, Klein, Pichon & Cleeremans · 2012
Behavioral Priming: It's All in the Mind, but Whose Mind?
replication experiments
What it supports: A famous word-to-walking-speed result was sensitive to replication conditions and experimenter expectations.
What it cannot establish: All priming is false, or all behaviour is unconsciously controlled.
Supplied research review: Publisher full-text HTML; abstract and methods context inspected
9. Tversky & Kahneman · 1981
The framing of decisions and the psychology of choice
experiments
What it supports: Equivalent risky choices can elicit different preferences when framed differently.
What it cannot establish: Every framing example is the same paradigm or always changes a decision.
Supplied research review: Publisher/indexed abstract and bibliographic record
10. Epley & Gilovich · 2006
The anchoring-and-adjustment heuristic: Why the adjustments are insufficient
experiments
What it supports: Examines adjustment away from initial anchors.
What it cannot establish: Every anchoring effect has one mechanism or every first number controls a decision.
Supplied research review: Indexed abstract and bibliographic record
11. Tversky & Kahneman · 1973
Availability: A heuristic for judging frequency and probability
experiments
What it supports: Ease of bringing examples to mind can enter frequency and probability judgments.
What it cannot establish: Seeing several reports establishes an actual increase in underlying events.
Supplied research review: Publisher abstract and bibliographic record
12. Weaver, Garcia, Schwarz & Miller · 2007
Inferring the popularity of an opinion from its familiarity: A repetitive voice can sound like a chorus
experiments
What it supports: Repeated opinions from one source can affect estimates of group support.
What it cannot establish: Repeated reports are independent evidence, or all consensus is manufactured.
Supplied research review: Indexed abstract and author-institution record
13. Muchnik, Aral & Taylor · 2013
Social Influence Bias: A Randomized Experiment
randomised online field experiment
What it supports: Initial rating signals affected subsequent online voting.
What it cannot establish: A like is belief, endorsement, expertise or evidence of truth.
Supplied research review: Publisher abstract and author research account
14. Hart and colleagues · 2009
Feeling validated versus being correct: A meta-analysis of selective exposure to information
meta-analysis
What it supports: People often prefer congenial information, with important motivations and situational moderators.
What it cannot establish: People never seek disagreement, or current feeds necessarily change their beliefs.
Supplied research review: Indexed abstract and research text surfaced
16. Green & Brock · 2000
The role of transportation in the persuasiveness of public narratives
experiments
What it supports: Narrative absorption was associated with, and experimentally related to, story-consistent responses.
What it cannot establish: Storytelling switches off critical thought in everyone.
Supplied research review: Indexed and author-institution abstracts
17. Ecker and colleagues · 2022
The psychological drivers of misinformation belief and its resistance to correction
research review
What it supports: Overview of belief formation, correction and persistence of misinformation.
What it cannot establish: Every correction backfires or a simple correction always erases prior influence.
Supplied research review: Publisher abstract, preview and reference list; full review not read
18. Begg, Anas & Farinacci · 1992
Dissociation of processes in belief: Source recollection, statement familiarity, and the illusion of truth
experiments
What it supports: Separates memory for a source from familiarity with a statement.
What it cannot establish: All source information is always forgotten.
Supplied research review: Publisher bibliographic record and author-uploaded abstract
23. Wood, Perunovic & Lee · 2009
Positive Self-Statements: Power for Some, Peril for Others
survey and experiments
What it supports: A positive-statement exercise had different short-term effects depending on participants' self-esteem.
What it cannot establish: All affirmations help; all affirmations harm; thoughts become literal facts.
Supplied research review: Publisher abstract read directly
24. Kross and colleagues · 2014
Self-talk as a regulatory mechanism: How you do it matters
experimental series
What it supports: Changing self-reference affected distancing and regulation under social stress.
What it cannot establish: Language writes identity into an unquestioning unconscious.
Supplied research review: Author-hosted full PDF; abstract and early methods inspected; page 1 visually checked
25. Hatzigeorgiadis, Zourbanos, Galanis & Theodorakis · 2011
Self-Talk and Sports Performance: A Meta-Analysis
meta-analysis
What it supports: Task-oriented self-talk interventions can help sport performance; task and training matter.
What it cannot establish: An all-purpose law of positive thinking or success without practice.
Supplied research review: Publisher abstract read directly
26. Epton, Harris, Kane, van Koningsbruggen & Sheeran · 2015
The impact of self-affirmation on health-behavior change: A meta-analysis
meta-analysis
What it supports: Values/attributes/relationships reflection showed small benefits on health-message and behaviour outcomes.
What it cannot establish: This is identical to repeating grandiose positive slogans or a medical treatment.
Supplied research review: Indexed abstract read directly
27. Zhang, Chen, Hu & Wang · 2025
The impact of self-affirmation interventions on well-being: A meta-analysis
meta-analysis
What it supports: Different self-affirmation interventions showed small average benefits on several well-being measures.
What it cannot establish: Universal benefits, automatic truth acceptance, or an effect of every popular mantra.
Supplied research review: Indexed abstract read directly; online publication date 27 October 2025
31. Webb & Sheeran · 2006
Does changing behavioral intentions engender behavior change? A meta-analysis of the experimental evidence
meta-analysis
What it supports: Changing intentions did not translate one-for-one into changes in behaviour.
What it cannot establish: Feeling, belief, intention and action are one outcome.
Supplied research review: Indexed and author-institution abstracts
32. Pennycook and colleagues · 2021
Shifting attention to accuracy can reduce misinformation online
experiments and field test
What it supports: Accuracy prompts improved sharing discernment in tested settings.
What it cannot establish: One pause inoculates someone permanently, or sharing proves belief.
Supplied research review: Publisher/indexed abstracts and research record
33. Roozenbeek and colleagues · 2022
Psychological inoculation improves resilience against misinformation on social media
preregistered experiments
What it supports: Brief explanations of manipulation techniques improved measured discernment outcomes.
What it cannot establish: The untested Lens article or umbrella activity has established protective efficacy.
Supplied research review: Publisher/indexed and author-institution abstracts
42. Shechter & Klauer · 2025
Rethinking Knowledge’s Impact on the Illusory Truth Effect
model comparison, reanalysis and experiments
What it supports: A reanalysis and two new experiments revisit the relative roles of knowledge and fluency; the new data supported a knowledge-precedence model.
What it cannot establish: Knowledge always prevents repetition effects, or the earlier experiments never found them.
Added for this candidate: indexed abstract read directly on 30 September 2026; full paper not reviewed.
AU_FEED_01 · Prime Minister of Australia · 2026-09-08
My Feed, My Way
What it supports: Draft Digital Duty of Care proposal and feed-choice description.
Where it stops: Government description; attribute evaluative rationale.
TRACK_01 · Google Ad Manager Help · current
Data feedback loop
What it supports: Personalised ads may use historical activity, visits, location, demographics and inferred interests.
Where it stops: Platform description; settings/practices vary.
Possible effects · We cannot say how likely
What this could change
Distinguishing an encounter from independent evidence can prompt readers to pause before repeating unverified claims, while understanding feed mechanics can clarify why familiar posts recur without broad consensus. The evidence does not measure how many readers will permanently alter their daily sharing or reading habits.
Repetition increases subjective truth ratings across laboratory cohorts even when prior knowledge is available.
Explaining how repetition mimics corroboration provides a cognitive framework for readers to inspect the originating record before passing on a claim.
Where the connection stopsThe research establishes experimental truth-rating shifts, not individual behaviour across a live social network.
What this depends on—and other possibilities
This depends on
- Readers retain the distinction between repetition and verification during live browsing.
- Metacognitive prompting influences informal conversational sharing.
Other explanations
- Affective reactions and existing group identities may override source monitoring.
- Speed of reading on mobile feeds may limit deliberate reflection.
Recommendation systems select and rank content based on interaction cues and profiling rather than factual verification.
Understanding that feeds select material using signals could help readers avoid mistaking widespread feed presence for universal agreement.
Where the connection stopsThe institutional records document proposed feed rules and tracking practices, not a proven shift in public skepticism.
What this depends on—and other possibilities
This depends on
- Readers who understand recommendation mechanics are less likely to assume a trending topic represents majority opinion.
- Feed transparency rules provide actionable controls for users.
Other explanations
- Users may feel resigned to algorithmic curation regardless of technical literacy.
- Platforms may continually adjust recommendation weights, changing exposure patterns.
How different interpretations could affect what happens next
How people may respond
How the story itself could change what happens
How the difference between repetition and independent corroboration is interpreted could change sharing habits, feed scepticism and inquiry into familiar claims.
What the evidence does not showCognitive fluency effects and recommendation mechanics are documented; live behavioural shifts across diverse social networks are not measured.
Repetition is recognized as separate from evidence
- How it is told
The investigation demonstrates how multiple exposures mimic confirmation even for inaccurate claims.
- What people may take from it
Readers may pause to verify the originating source before adopting or amplifying familiar posts.
- Where attention could turn
Attention moves from how familiar a claim feels to whether primary evidence exists.
- What people may do
Readers, researchers and online communities: Apply source-checking habits and question repetitive narrative loops.
- What could change
Unverified claims face greater scrutiny in peer sharing and personal decision-making.
What we know has changedWe have not established that this possible change has happened.
What this does not showCognitive awareness does not eliminate emotional or social drivers of sharing.
Why we are cautious
Why we cannot tell yet
This is the first time Lens has mapped this path. We have no later evidence showing whether it is happening more, less or about the same.
Signs that would support this path
- Adoption of source-checking interventions and downstream sharing patterns in communication studies.
Signs that would weaken it
- Emotional or partisan cues consistently bypass source-monitoring checks.
- Automated bot amplification drowns out human verification pauses.
This depends on
- Readers retain the distinction between repetition and verification during live browsing.
- Metacognitive prompting influences informal conversational sharing.
- Readers who understand recommendation mechanics are less likely to assume a trending topic represents majority opinion.
- Feed transparency rules provide actionable controls for users.
Volume and algorithmic repetition dominate nuanced correction
- How it is told
High-velocity feeds continuously circulate engagement-optimised claims.
- What people may take from it
Audiences may treat ubiquity as consensus, assuming widespread presence equates to factual truth.
- Where attention could turn
Attention concentrates on trending volume while corrections remain unseen.
- What people may do
Platforms, audiences and content creators: Continue circulating familiar narratives without inspecting primary documentation.
- What could change
Familiar claims maintain high subjective credibility regardless of published corrections.
What we know has changedWe have not established that this possible change has happened.
What this does not showVolume influences perception, but does not prevent determined readers from verifying facts.
Why we are cautious
Why we cannot tell yet
This is the first time Lens has mapped this path. We have no later evidence showing whether it is happening more, less or about the same.
Signs that would support this path
- Persistence of debunked claims in longitudinal belief tracking surveys.
Signs that would weaken it
- Platform architecture changes mandate clear provenance and source indicators.
- Regulatory standards require verifiable chronological feed alternatives.
This depends on
- Readers retain the distinction between repetition and verification during live browsing.
- Metacognitive prompting influences informal conversational sharing.
- Readers who understand recommendation mechanics are less likely to assume a trending topic represents majority opinion.
- Feed transparency rules provide actionable controls for users.
Two ways this could develop
Metacognitive friction reduces reflexive amplification
If readers apply source-checking habits and distinguish multiple sightings from multiple independent witnesses
Then unsubstantiated claims face higher verification barriers in informal peer sharing and personal decision-making.
What to watch—and what would weaken it
- Adoption of source-checking interventions and downstream sharing patterns in communication studies.Academic media literacy and behavioural research.
Would weaken this: Emotional or partisan cues consistently bypass source-monitoring checks. Automated bot amplification drowns out human verification pauses.
Scope: Reader comprehension and voluntary verification habits. Horizon: Ongoing media consumption.
Repetition volume overpowers downstream nuance
If high-volume repetition in curated feeds continues to create fluency effects that outpace deliberate verification
Then familiar false claims retain high perceived credibility despite accessible evidence and institutional corrections.
What to watch—and what would weaken it
- Persistence of debunked claims in longitudinal belief tracking surveys.Public opinion monitoring and misinformation tracking.
Would weaken this: Platform architecture changes mandate clear provenance and source indicators. Regulatory standards require verifiable chronological feed alternatives.
Scope: Ecosystem-wide information dynamics. Horizon: Medium-term platform and regulatory cycles.
How do we know?Inspect the evidence and its limits
Evidence used in this assessment
Repetition increased perceived truth on average; effects varied considerably.
Open evidence ↗Repeated opinions from one source can affect estimates of group support.
Open evidence ↗Changing intentions did not translate one-for-one into changes in behaviour.
Open evidence ↗Draft Digital Duty of Care proposal and feed-choice description.
Open evidence ↗Personalised ads may use historical activity, visits, location, demographics and inferred interests.
Open evidence ↗What could change this assessment?
- New pre-registered field trials testing real-world social media sharing interventions.
- Passage and enforcement of Australian statutory feed-choice regulations.
- Platform policy updates regarding algorithmic provenance labelling.
Free Lens updates
Understand the story. Get the next Lens.
Our free weekly email brings you the latest Lens reporting and what changed.