Lightnews — Scholar-powered news

Xiaoli Nan

@drxnan.bsky.social

90 followers 130 following 4 posts

Professor of Communication Science @UMD | persuasion, health communication, misinformation, AI, media effects

Videos

Psychology 26%

Sociology 15%

Posts Media Videos Starter Packs

Xiaoli Nan @drxnan.bsky.social · 9d

🚨New research from our lab! Generative AI and misinformation: a scoping review of the role of generative AI in the generation, detection, mitigation, and impact of misinformation link.springer.com/article/10.1...

Generative AI and misinformation: a scoping review of the role of generative AI in the generation, detection, mitigation, and impact of misinformation - AI & SOCIETY

The rapid advancement of generative artificial intelligence (AI) has introduced both opportunities and challenges in the fight against misinformation. This scoping review synthesizes recent empirical studies to explore the dual role of generative AI—particularly large language models (LLMs)—in the generation, detection, mitigation, and impact of misinformation. Analyzing 24 empirical studies, our review suggests that LLMs can generate highly convincing misinformation, often exploiting cognitive biases and ideological leanings of the audiences, while also demonstrating the ability to detect false claims and enhance users’ resistance to misinformation. Mitigation efforts show mixed results, with personalized corrections proving effective but safeguards inconsistently applied. Additionally, exposure to AI-generated misinformation was found to reduce trust and influence decision-making. This review underscores the need for standardized evaluation metrics, interdisciplinary collaboration, and stronger regulatory measures to ensure the responsible use of generative AI in the information ecosystem.

link.springer.com

Reposted by Xiaoli Nan

Timothy Caulfield @caulfieldtim.bsky.social · 15d

WHO joins global health leaders rejecting U.S. acetaminophen warnings www.washingtonpost.com/health/2025/...

@who.int: “no conclusive scientific evidence confirming a possible link.”

"Public health leaders and organizations around the world have been sounding off..."

U.S. become Dark Age Nation

WHO joins global health leaders rejecting U.S. acetaminophen warnings

The World Health Organization is the latest to criticize claims from President Donald Trump and Robert F. Kennedy Jr. linking autism to Tylenol use during pregnancy.

www.washingtonpost.com

3 25 73

Reposted by Janet K. Swim, William C. Clark, Xiaoli Nan , and 1 more Janet K. Swim, William C. Clark, Xiaoli Nan, Meru Sheel

Kai Kupferschmidt @kakape.bsky.social · Dec 14

Sadly, it’s a good time to once again share this amazing infographic that we ran at @science.org more than 7 years ago
🧪 #IDsky
www.science.org/content/arti...

Here's the visual proof of why vaccines do more good than harm

See year by year how vaccines beat back nine dangerous infectious diseases

www.science.org

230 3.7K 10K

Reposted by Xiaoli Nan

Guy Grossman @guygrossman.bsky.social · May 12

We are going to miss the days when the GOP official position on vaccines was that it’s “personal choice “

4 12 35

Reposted by Xiaoli Nan

Matthew Facciani @matthewfacciani.bsky.social · Apr 14

A new study tested five interventions—including value-based messaging and showcasing conservative scientists—but none increased conservatives’ trust in science, suggesting these attitudes are stable and resistant to brief messaging.
www.nature.com/articles/s41... #ScienceSky #SciComm

Political ideology and trust in scientists in the USA - Nature Human Behaviour

In this Registered Report, Gligorić et al. find that liberals in the USA tend to have higher trust in most scientists compared with conservatives. However, they find no evidence that a series of inter...

www.nature.com

4 17 51

Xiaoli Nan @drxnan.bsky.social · Mar 30

Just assume we authors will have to deal with not only AI reviewers but also AI editors..and vice versa…

Xiaoli Nan @drxnan.bsky.social · Mar 30

I agree; if AI is used to polish and organize thoughts that’s fine. However it was clear several comments were directly generated by AI focusing on obscure technical issues that a generalist editor is unlikely to know.

1 1

Xiaoli Nan @drxnan.bsky.social · Mar 30

We once got a desk reject from a top comm journal and the editor’s comments are clearly AI generated. Is that ethical?

1 2

Reposted by Xiaoli Nan

Rakoen Maertens @rakoenmaertens.bsky.social · Mar 11

|| New Publication (Nature Communications)

Psychological Booster Shots Targeting Memory Increase Long-Term Resistance Against Misinformation
doi.org/10.1038/s414...

Press Release
www.ox.ac.uk/news/2025-03...

Key Findings
* Inoculation works
* Effects dissipate
* Booster shots remedy this

(1/10)

5 44 99

Reposted by Xiaoli Nan

Gary Marcus @garymarcus.bsky.social · Feb 16

Elon Musk’s terrifying vision for AI; short, urgent new piece at Marcus on AI.

I can’t remain silent - and hope you won’t, either.

open.substack.com/pub/garymarc...

Elon Musk’s terrifying vision for AI

All your thoughts belong to him

open.substack.com

10 68 200

Reposted by Xiaoli Nan

Stephan Lewandowsky @lewan.bsky.social · Jan 26

Juste came across this interesting paper: repetition does not just cause "illusory truth" but also "illusory implications" -- people make inferences on the basis of familiar (illusory) information. @lkfazio.bsky.social @profsanderlinden.bsky.social @ulliecker.bsky.social doi.org/10.1098/rsos...

4 20 70

Reposted by Xiaoli Nan, John R. Kerr

Sander van der Linden @profsanderlinden.bsky.social · Jan 26

The spread of misinformation and disinformation top the WEF short-term global risk report for 2025. www.weforum.org/publications...

7 58 110

Reposted by Xiaoli Nan

Matthew Facciani @matthewfacciani.bsky.social · Jan 26

Instead of directly correcting false claims, “bypassing” offers an alternative strategy: focus on positive, truthful statements about the same topic. New research suggests it can be very effective and I cover it in more detail here!
substack.com/home/post/p-... #MisinfoResearch

"Bypassing" is a new way to counter misinformation, but how does it work?

A new study found that "bypassing" can counter misinformation without direct confrontation.

substack.com

17 110 300

Reposted by Xiaoli Nan

Kim Doell @kimdoell.bsky.social · Jan 17

Our paper "Addressing climate change with behavioral science: A global intervention tournament in 63 countries" reached 100 citations today 🤓 (within a year of being published)! Huge thanks to our 258 co-authors.
www.science.org/doi/10.1126/...

Addressing climate change with behavioral science: A global intervention tournament in 63 countries

Climate interventions increase beliefs, policy support, and willingness to share information but not higher effort action.

www.science.org

7 24

Reposted by Xiaoli Nan

Eric Topol @erictopol.bsky.social · Jan 16

This week's cover and editorial @thelancet.bsky.social on mis- and disinformation's impact on public health
www.thelancet.com/journals/lan...

6 95 230

Reposted by Xiaoli Nan

Richard Huskey @richardhuskey.bsky.social · Jan 2

When it comes to statistical power, how is Communication Science doing? Sun, Shen, Pan, and Quan’s article: “Toward a More Powerful Experimental Communication Science: An Assessment of Two Decades’ Research (2001–2023)” gives an answer. In short, not so good: doi.org/10.1177/0093...

A bar chart displaying the reporting of statistical power analysis by year from 2001 to 2023. The Y-axis shows the percentage of studies per category within each year, ranging from 0% to 75%. The X-axis represents years from 2001 to 2023. The chart is divided into five categories listed on the right: “A Priori,” “Sensitivity,” “Observed,” “Mention Only,” and “No Mention.”

A Priori: Reporting is sparse before 2010 but increases steadily after 2017, peaking near 25% in recent years.

Sensitivity: Reporting is minimal and sporadic across all years.

Observed: Reporting is consistent but moderate, fluctuating around 10–25%.

Mention Only: Reporting is infrequent and fluctuates around 5–10%.

No Mention: Initially dominant but decreases over time, particularly after 2010.

The chart highlights a trend toward increased A Priori power reporting and a decline in studies making no mention of power analysis, indicating evolving research practices over time.

Table 2 summarizes the results from power analyses by study design, showing statistical power (median and range) and the percentage of studies at different power levels. The table includes data for four designs: Within-subject (k = 41), Two-group (k = 98), Multi-group (k = 93), and 2 × 2 Factorial (k = 184).

The table shows that within-subject designs generally achieve higher statistical power, especially at larger effect sizes, while two-group and multi-group designs often fall below acceptable power levels, particularly for small effects. 2 × 2 factorial designs vary, with reversal effects showing higher power than attenuation effects. Overall, achieving adequate power remains a challenge in many designs, highlighting the need for careful study planning.

1 5 14

Reposted by Robert Böhm, Aleksandra Cichocka, Xiaoli Nan

Jay Van Bavel, PhD @jayvanbavel.bsky.social · Jan 2

We have a new paper explaining all the ways you can use natural language processing to analyze text data in @natrevpsych.bsky.social

We provide user friendly recommendations for using NLP to ensure rigour and reproducibility

Here is a free link: www.nature.com/articles/s44...

4 53 150

Reposted by Emily K. Vraga, Kai Sassenberg, Sebastián Valenzuela , and 2 more Emily K. Vraga, Kai Sassenberg, Sebastián Valenzuela, Xiaoli Nan, Jana Lasser

Sander van der Linden @profsanderlinden.bsky.social · Dec 16

In this new article in American Psychologist we respond to critics in detail and clarify two key points for the field;

(1) The prevalence of misinformation in society is substantial when properly defined.

(2) Misinformation causally impacts attitudes and behaviors.

psycnet.apa.org/fulltext/202...

44 280 890

Reposted by Xiaoli Nan

Helen Branswell 🇨🇦 @helenbranswell.bsky.social · Dec 26

What lies ahead in global health in 2025? I can't tell you if #H5N1 #birdflu is going to trigger a pandemic, but I can tell you a few things I'm going to be watching in the new year. www.statnews.com/2024/12/26/g...

Three issues to watch in global health in 2025

Three issues to watch in global health in 2025: H5N1 bird flu, mpox, and U.S. influence on the global stage.

www.statnews.com

12 110 240

Reposted by Joshua A. Tucker, Xiaoli Nan

Jay Van Bavel, PhD @jayvanbavel.bsky.social · Dec 23

This is consistent with a review of 747 COVID papers we published in @nature.com finding that evidence for social science and behavior claims during COVID as very high (16/18 claims were supported) and this research was highly rigorrous (average N= 16,848) www.nature.com/articles/s41...

A synthesis of evidence for policy from behavioural science during COVID-19 - Nature

Evaluation of evidence generated to test 19 proposed policy recommendations and guidance for the future.

www.nature.com

3 9 32

Reposted by Xiaoli Nan

Brian Nosek @briannosek.bsky.social · Dec 23

New in NHB: An appetizer from the SCORE project. We replicated 29 social science COVID papers. Beginners and more experienced researchers forecasted the replication outcomes following a structured elicitation process and were similarly successful at prediction, ~65%.

www.nature.com/articles/s41...

2 19 48

Reposted by Xiaoli Nan

Carl T. Bergstrom @carlbergstrom.com · Dec 22

We wrote about this at some length in our paper "Misinformation in and about science".

www.pnas.org/doi/10.1073/...

$In addition, researchers commonly misstate or overstate the implications of their work (35). In concert with researchers, university press offices play a particularly important role in communicating science—but too frequently do so in ways that prioritize web traffic over accuracy. Sometimes spin is carried over from the journal article itself (36); other times, it is added in the press release. A biomedical report might omit important caveats, draw inappropriate extrapolations from mouse models, and exaggerate prescriptive implications. One analysis found that nearly a third of 525 papers in top obesity or nutrition journals make inappropriate causal claims in their abstracts or titles (37); in another study, roughly the same fraction of health-related papers widely shared on social media used inappropriately strong causal language (38). Some fields may be more prone to hype than others. A new result on the geometry of Banach spaces may be more difficult to hype than a bioweapon claim, but we surmise that most fields are susceptible. Much of this truth bending may be unnecessary. Most studies (39–41) fail to find an association between exaggeration and uptake by the news media. Admittedly, selection bias may play a role: perhaps the stories that are not exaggerated are those that do not need to be. In any case, high-quality press releases appear to drive higher-quality news stories on that research (42, 43).$

1 10 58

Reposted by Robert West, Emily K. Vraga, David A. Broniatowski , and 2 more Robert West, Emily K. Vraga, David A. Broniatowski, Xiaoli Nan, John R. Kerr

Lisa Fazio @lkfazio.bsky.social · Dec 19

Out now - National Academies consensus report on Understanding and Addressing Misinformation About Science 🧪

It was a privilege to serve as one of the 15 committee members from a wide range of scientific disciplines who put this report together. Quick 🧵1/

www.nationalacademies.org/our-work/und...

Cover of the report - it's in a blue/purple color scheme and shows an abstract illustration of a molecule and the title of the report "Understanding and Addressing Misinformation About Science Consensus Study Report"

17 170 440

Reposted by Joshua A. Tucker, William J. Brady, Christian Breunig , and 8 more Joshua A. Tucker, William J. Brady, Christian Breunig, Yamil Velez, Xiaoli Nan, Paul Goldsmith-Pinkham, Juanjo Medina, Hannes Rusch, Kohei Watanabe, Gregory Eady, Stephen A. Meserve

Christopher Barrie @cbarrie.bsky.social · Dec 17

Pleased to share the latest version of my paper with Arthur Spirling and @lexipalmer.bsky.social on replication using LMs

We show:

1. current applications of LMs in political science research *don't* meet basic standards of reproducibility...

18 170 440

Reposted by Xiaoli Nan

Matthew Facciani @matthewfacciani.bsky.social · Dec 17

Conspiracy theory believers who interacted with an AI chatbot that asked them to reflect on the uncertainties of their beliefs had significantly less conviction in their conspiracy theory beliefs.
misinforeview.hks.harvard.edu/article/usin... #MisinfoResearch

Using an AI-powered “street epistemologist” chatbot and reflection tasks to diminish conspiracy theory beliefs | HKS Misinformation Review

Social scientists, journalists, and policymakers are increasingly interested in methods to mitigate or reverse the public’s beliefs in conspiracy theories, particularly those associated with negative ...

misinforeview.hks.harvard.edu

3 5 18

Reposted by Aleksandra Cichocka, Xiaoli Nan

Mikey Biddlestone @mikeybiddlestone.bsky.social · Dec 17

🚨Our meta-analysis of the motives associated with conspiracy beliefs has been accepted at Psychological Bulletin!🚨“Reasons to believe: A systematic review and meta-analytic synthesis of the motives associated with conspiracy beliefs” osf.io/preprints/ps... 🧵👇 1/16

Abstract for our accepted pre-print of Reasons to believe: A systematic review and meta-analytic synthesis of the motives associated with conspiracy beliefs

3 21 51

Reposted by Robb Willer, Pablo Brañas‐Garza, Xiaoli Nan , and 1 more Robb Willer, Pablo Brañas‐Garza, Xiaoli Nan, Jessica Gall Myrick

David Rand @dgrand.bsky.social · Dec 16

🚨In Science🚨
Conspiracy beliefs famously resist correction, ya?
WRONG: We show brief convos w GPT4 reduce conspiracy beliefs by ~20%!
-Lasts over 2mo
-Works on entrenched beliefs
-Tailored AI response rebuts specific evidence offered by believers
www.science.org/doi/10.1126/...
1/
[X->BSky repost]

19 90 280

Reposted by Xiaoli Nan

koenfucius @koenfucius.bsky.social · Dec 15

Research by @dbroockman.bsky.social et al suggests
•experienced political practitioners and laypeople perform barely better than chance at predicting which messages will be most persuasive
•practitioners’ predictions are not meaningfully better than laypeople’s: https://buff.ly/3OU7rrB

1 9 19

Reposted by Xiaoli Nan

Matt Grossmann @mattgrossmann.bsky.social · Dec 14

people are good at discerning true from false information but partisan bias in responses to true and false information is pervasive & strong; skepticism against belief-incongruent true information is much more pronounced than gullibility to false info
journals.sagepub.com/doi/full/10....

Sage Journals: Discover world-class research

Subscription and open access journals from Sage, the world's leading independent academic publisher.

journals.sagepub.com

8 21