Sacha Altay
@sachaltay.bsky.social
🔎 Misinformation, social media & the news 🗞️
🇨🇭 Postdoc at the University of Zurich, previously Reuters Institute & ENS 🇫🇷
🇨🇭 Postdoc at the University of Zurich, previously Reuters Institute & ENS 🇫🇷
Reposted by Sacha Altay
Ok, just wow. If the content of this article is right, this is depressing. We're slowly reaching the point where ~100% of what I was taught in Social Psych was either innocently wrong or plainly frauded
onlinelibrary.wiley.com/doi/abs/10.1...
onlinelibrary.wiley.com/doi/abs/10.1...
Debunking “When Prophecy Fails”
In 1954, Dorothy Martin predicted an apocalyptic flood and promised her followers rescue by flying saucers. When neither arrived, she recanted, her group dissolved, and efforts to proselytize ceased....
onlinelibrary.wiley.com
November 5, 2025 at 11:09 PM
Ok, just wow. If the content of this article is right, this is depressing. We're slowly reaching the point where ~100% of what I was taught in Social Psych was either innocently wrong or plainly frauded
onlinelibrary.wiley.com/doi/abs/10.1...
onlinelibrary.wiley.com/doi/abs/10.1...
Reposted by Sacha Altay
We have a new preprint: osf.io/preprints/so...
What have we learned about social media - the constantly moving target of empirical research - over the past decade?
What have we learned about social media - the constantly moving target of empirical research - over the past decade?
October 30, 2025 at 10:53 AM
We have a new preprint: osf.io/preprints/so...
What have we learned about social media - the constantly moving target of empirical research - over the past decade?
What have we learned about social media - the constantly moving target of empirical research - over the past decade?
Reposted by Sacha Altay
Happy to share our new preprint “Content-based detection of misinformation expands its scope across politicians and platforms.” We analyzed misinformation at the text level in posts by German politicians on Facebook, Instagram, X, and TikTok.
osf.io/preprints/so... 🧵1/8
osf.io/preprints/so... 🧵1/8
OSF
osf.io
November 3, 2025 at 10:13 AM
Happy to share our new preprint “Content-based detection of misinformation expands its scope across politicians and platforms.” We analyzed misinformation at the text level in posts by German politicians on Facebook, Instagram, X, and TikTok.
osf.io/preprints/so... 🧵1/8
osf.io/preprints/so... 🧵1/8
Reposted by Sacha Altay
See also Peterson and Iyengar onlinelibrary.wiley.com/doi/full/10.... - "our findings support the motivated reasoning interpretation of misinformation; partisans seek out information with congenial slant and sincerely adopt inaccurate beliefs that cast their party in a favorable light."
After a huge post-election flip in economic perceptions, I thought Democrats and Republicans might be lying to pollsters to send a partisan message — but I was wrong!
New in the Journal of Experimental Political Science (open access): doi.org/10.1017/XPS....
New in the Journal of Experimental Political Science (open access): doi.org/10.1017/XPS....
October 27, 2025 at 6:06 PM
See also Peterson and Iyengar onlinelibrary.wiley.com/doi/full/10.... - "our findings support the motivated reasoning interpretation of misinformation; partisans seek out information with congenial slant and sincerely adopt inaccurate beliefs that cast their party in a favorable light."
Reposted by Sacha Altay
What if people appreciate having an abundance of content and communication, more than they feel overloaded by it?
@annisch.bsky.social et al decided to have a look. Their results? "We found that appreciation for abundance was about twice as common as overload".
Paper: journalqd.org/article/view...
@annisch.bsky.social et al decided to have a look. Their results? "We found that appreciation for abundance was about twice as common as overload".
Paper: journalqd.org/article/view...
October 31, 2025 at 9:33 AM
What if people appreciate having an abundance of content and communication, more than they feel overloaded by it?
@annisch.bsky.social et al decided to have a look. Their results? "We found that appreciation for abundance was about twice as common as overload".
Paper: journalqd.org/article/view...
@annisch.bsky.social et al decided to have a look. Their results? "We found that appreciation for abundance was about twice as common as overload".
Paper: journalqd.org/article/view...
Reposted by Sacha Altay
NEW: In a recent Oxford-led study by @mmosleh.bsky.social, researchers analysed millions of social media posts containing links to news stories, across seven different social platforms.
Read the full research paper here: www.pnas.org/doi/10.1073/...
Read the full research paper here: www.pnas.org/doi/10.1073/...
October 31, 2025 at 1:24 PM
NEW: In a recent Oxford-led study by @mmosleh.bsky.social, researchers analysed millions of social media posts containing links to news stories, across seven different social platforms.
Read the full research paper here: www.pnas.org/doi/10.1073/...
Read the full research paper here: www.pnas.org/doi/10.1073/...
Reposted by Sacha Altay
"only about 5% of the variance in personality can be
predicted from digital footprints, and personality‐tailored messages show negligible effects on behavior... When design and evaluation flaws are controlled, the combined end‐to‐end effectiveness of psychological targeting approaches zero."
predicted from digital footprints, and personality‐tailored messages show negligible effects on behavior... When design and evaluation flaws are controlled, the combined end‐to‐end effectiveness of psychological targeting approaches zero."
The (In)Effectiveness of Psychological Targeting: A Meta‐Analytic Review
The use of psychological targeting—employing machine learning to predict consumer personality from digital footprints and subsequently tailoring persuasive messages—has emerged as a controversial yet...
onlinelibrary.wiley.com
October 31, 2025 at 2:55 PM
"only about 5% of the variance in personality can be
predicted from digital footprints, and personality‐tailored messages show negligible effects on behavior... When design and evaluation flaws are controlled, the combined end‐to‐end effectiveness of psychological targeting approaches zero."
predicted from digital footprints, and personality‐tailored messages show negligible effects on behavior... When design and evaluation flaws are controlled, the combined end‐to‐end effectiveness of psychological targeting approaches zero."
Reposted by Sacha Altay
🤖 𝐆𝐞𝐧𝐞𝐫𝐚𝐭𝐢𝐯𝐞 𝐀𝐈 𝐢𝐧 𝐭𝐡𝐞 𝟐𝟎𝟐𝟓 𝐃𝐮𝐭𝐜𝐡 𝐄𝐥𝐞𝐜𝐭𝐢𝐨𝐧: 𝐑𝐞𝐬𝐮𝐥𝐭𝐬 𝐟𝐫𝐨𝐦 𝐨𝐮𝐫 𝐂𝐚𝐦𝐩𝐀𝐈𝐠𝐧 𝐓𝐫𝐚𝐜𝐤𝐞𝐫 🇳🇱
Less than one week before the Dutch elections, @meinungsfuehrer.bsky.social & I looked at over 20k posts to reveal patterns in the usage of AI-generated visuals across Facebook, Instagram, TikTok and X. #tk2025
🧵Thread
Less than one week before the Dutch elections, @meinungsfuehrer.bsky.social & I looked at over 20k posts to reveal patterns in the usage of AI-generated visuals across Facebook, Instagram, TikTok and X. #tk2025
🧵Thread
October 23, 2025 at 9:36 AM
🤖 𝐆𝐞𝐧𝐞𝐫𝐚𝐭𝐢𝐯𝐞 𝐀𝐈 𝐢𝐧 𝐭𝐡𝐞 𝟐𝟎𝟐𝟓 𝐃𝐮𝐭𝐜𝐡 𝐄𝐥𝐞𝐜𝐭𝐢𝐨𝐧: 𝐑𝐞𝐬𝐮𝐥𝐭𝐬 𝐟𝐫𝐨𝐦 𝐨𝐮𝐫 𝐂𝐚𝐦𝐩𝐀𝐈𝐠𝐧 𝐓𝐫𝐚𝐜𝐤𝐞𝐫 🇳🇱
Less than one week before the Dutch elections, @meinungsfuehrer.bsky.social & I looked at over 20k posts to reveal patterns in the usage of AI-generated visuals across Facebook, Instagram, TikTok and X. #tk2025
🧵Thread
Less than one week before the Dutch elections, @meinungsfuehrer.bsky.social & I looked at over 20k posts to reveal patterns in the usage of AI-generated visuals across Facebook, Instagram, TikTok and X. #tk2025
🧵Thread
Reposted by Sacha Altay
New hobby:
Remaking article abstracts as movie trailers to expose hype and fearmongering.
Remaking article abstracts as movie trailers to expose hype and fearmongering.
October 20, 2025 at 10:22 AM
New hobby:
Remaking article abstracts as movie trailers to expose hype and fearmongering.
Remaking article abstracts as movie trailers to expose hype and fearmongering.
Reposted by Sacha Altay
We continue with our seminar series next Wednesday with Dr Sacha Altay @sachaltay.bsky.social from University of Zürich, talking about the problem of misinformation. Full details on our website (surl.li/mqtofs). See you there!
October 16, 2025 at 1:08 PM
We continue with our seminar series next Wednesday with Dr Sacha Altay @sachaltay.bsky.social from University of Zürich, talking about the problem of misinformation. Full details on our website (surl.li/mqtofs). See you there!
Super cool data on what people saw on their phone during the 2024 US elections:
"On an average day, the median individual saw the terms ‘Donald Trump’, ‘Joe Biden’, and ‘Kamala Harris’ on their phone for only 3 seconds."
www.guyaridor.net/files/digita...
"On an average day, the median individual saw the terms ‘Donald Trump’, ‘Joe Biden’, and ‘Kamala Harris’ on their phone for only 3 seconds."
www.guyaridor.net/files/digita...
www.guyaridor.net
October 14, 2025 at 9:46 AM
Super cool data on what people saw on their phone during the 2024 US elections:
"On an average day, the median individual saw the terms ‘Donald Trump’, ‘Joe Biden’, and ‘Kamala Harris’ on their phone for only 3 seconds."
www.guyaridor.net/files/digita...
"On an average day, the median individual saw the terms ‘Donald Trump’, ‘Joe Biden’, and ‘Kamala Harris’ on their phone for only 3 seconds."
www.guyaridor.net/files/digita...
Reposted by Sacha Altay
Climate skeptics actively engage in information avoidance. We need to bring them to the facts and let the facts do their work:
❝Using generative AI to increase sceptics’ engagement with climate science❞
(thread)
www.nature.com/articles/s41...
❝Using generative AI to increase sceptics’ engagement with climate science❞
(thread)
www.nature.com/articles/s41...
Using generative AI to increase sceptics’ engagement with climate science - Nature Climate Change
Climate sceptics tend to avoid climate information, making it even harder to reduce scepticism. This study shows that generative AI can enhance sceptics’ engagement with climate news by tailoring headlines to their existing perspective and shift their beliefs towards the scientific consensus.
www.nature.com
October 13, 2025 at 9:13 AM
Climate skeptics actively engage in information avoidance. We need to bring them to the facts and let the facts do their work:
❝Using generative AI to increase sceptics’ engagement with climate science❞
(thread)
www.nature.com/articles/s41...
❝Using generative AI to increase sceptics’ engagement with climate science❞
(thread)
www.nature.com/articles/s41...
Reposted by Sacha Altay
🗞️ 🤖 Weekend reading anyone? For the launch of @transformernews.ai as a standalone publication, they invited me to contribute a piece on what persuasive AI might mean for democracy and elections.
Here’s the result…
buff.ly/OJsNmpK
Here’s the result…
buff.ly/OJsNmpK
AI is persuasive, but that’s not the real problem for democracy
Opinion: Felix M Simon argues that AI is unlikely to significantly shape election results in the near future, but warns that it could damage democracy through a steady erosion of institutional trust.
www.transformernews.ai
October 3, 2025 at 5:19 PM
🗞️ 🤖 Weekend reading anyone? For the launch of @transformernews.ai as a standalone publication, they invited me to contribute a piece on what persuasive AI might mean for democracy and elections.
Here’s the result…
buff.ly/OJsNmpK
Here’s the result…
buff.ly/OJsNmpK
Reposted by Sacha Altay
WhatsApp deactivation experimental results in Brazil, India, and South Africa mirror those from prior social media studies: big changes in content exposure -> minimal changes in attitudes, decreases in political knowledge, increases in subjective wellbeing www.venturatiago.com/talk/vmtn_wh...
September 29, 2025 at 12:08 AM
WhatsApp deactivation experimental results in Brazil, India, and South Africa mirror those from prior social media studies: big changes in content exposure -> minimal changes in attitudes, decreases in political knowledge, increases in subjective wellbeing www.venturatiago.com/talk/vmtn_wh...
Reposted by Sacha Altay
NEW PAPER!!
We study how the "AI slop" era could actually boost demand for credible news.
In an experiment with thousands of Süddeutsche Zeitung readers, we found that AI misinformation made people *trust news less*, but *read it more*. 🧵
We study how the "AI slop" era could actually boost demand for credible news.
In an experiment with thousands of Süddeutsche Zeitung readers, we found that AI misinformation made people *trust news less*, but *read it more*. 🧵
August 11, 2025 at 11:44 AM
NEW PAPER!!
We study how the "AI slop" era could actually boost demand for credible news.
In an experiment with thousands of Süddeutsche Zeitung readers, we found that AI misinformation made people *trust news less*, but *read it more*. 🧵
We study how the "AI slop" era could actually boost demand for credible news.
In an experiment with thousands of Süddeutsche Zeitung readers, we found that AI misinformation made people *trust news less*, but *read it more*. 🧵
Reposted by Sacha Altay
new working paper: osf.io/7v4zj
tl;dr: taking the "marketplace of ideas" metaphor seriously requires taking the nuances of actually-existing markets seriously, which means thinking about tradeoffs between transaction costs and conformity costs. applied to misinfo and ideological belief systems.
tl;dr: taking the "marketplace of ideas" metaphor seriously requires taking the nuances of actually-existing markets seriously, which means thinking about tradeoffs between transaction costs and conformity costs. applied to misinfo and ideological belief systems.
August 4, 2025 at 2:59 PM
new working paper: osf.io/7v4zj
tl;dr: taking the "marketplace of ideas" metaphor seriously requires taking the nuances of actually-existing markets seriously, which means thinking about tradeoffs between transaction costs and conformity costs. applied to misinfo and ideological belief systems.
tl;dr: taking the "marketplace of ideas" metaphor seriously requires taking the nuances of actually-existing markets seriously, which means thinking about tradeoffs between transaction costs and conformity costs. applied to misinfo and ideological belief systems.
Reposted by Sacha Altay
From ancient Greece to the Arabic golden age, scholars have been driven by their curiosity to investigate astronomy, history, philosophy, and sundry other disciplines. Is there a structure to that curiosity? Are astronomers as likely to also be historians or to also be philosophers?
August 11, 2025 at 12:05 PM
From ancient Greece to the Arabic golden age, scholars have been driven by their curiosity to investigate astronomy, history, philosophy, and sundry other disciplines. Is there a structure to that curiosity? Are astronomers as likely to also be historians or to also be philosophers?
Reposted by Sacha Altay
New paper with @kirstenaeddy.bsky.social in Journalism on whether people across 45 countries think news coverage of people their age, their gender, and with their political views is fair or unfair.
We outline some interesting demographic differences (more details in 🧵)...
doi.org/10.1177/1464...
We outline some interesting demographic differences (more details in 🧵)...
doi.org/10.1177/1464...
August 11, 2025 at 11:33 AM
New paper with @kirstenaeddy.bsky.social in Journalism on whether people across 45 countries think news coverage of people their age, their gender, and with their political views is fair or unfair.
We outline some interesting demographic differences (more details in 🧵)...
doi.org/10.1177/1464...
We outline some interesting demographic differences (more details in 🧵)...
doi.org/10.1177/1464...
This is good news for debriefings in misinformation research:
Debriefings that include fact-checks of the false claims used in the experiment reduce false beliefs, improve attitudes towards the study, and show no signs of negative effects.
👉 osf.io/preprints/ps...
Feedback much welcomed :)
Debriefings that include fact-checks of the false claims used in the experiment reduce false beliefs, improve attitudes towards the study, and show no signs of negative effects.
👉 osf.io/preprints/ps...
Feedback much welcomed :)
July 29, 2025 at 7:59 AM
This is good news for debriefings in misinformation research:
Debriefings that include fact-checks of the false claims used in the experiment reduce false beliefs, improve attitudes towards the study, and show no signs of negative effects.
👉 osf.io/preprints/ps...
Feedback much welcomed :)
Debriefings that include fact-checks of the false claims used in the experiment reduce false beliefs, improve attitudes towards the study, and show no signs of negative effects.
👉 osf.io/preprints/ps...
Feedback much welcomed :)
Reposted by Sacha Altay
What about debriefings in Misinformation Research?
✅ Only specific debriefings reduce belief in false info
⚠️ Both slightly reduce belief in true info
🧠 Only specific boosts perceived learning
🟢 Only general reduces manipulation & frustration
🔍 Both increase transparency
osf.io/preprints/ps...
✅ Only specific debriefings reduce belief in false info
⚠️ Both slightly reduce belief in true info
🧠 Only specific boosts perceived learning
🟢 Only general reduces manipulation & frustration
🔍 Both increase transparency
osf.io/preprints/ps...
July 28, 2025 at 6:35 PM
What about debriefings in Misinformation Research?
✅ Only specific debriefings reduce belief in false info
⚠️ Both slightly reduce belief in true info
🧠 Only specific boosts perceived learning
🟢 Only general reduces manipulation & frustration
🔍 Both increase transparency
osf.io/preprints/ps...
✅ Only specific debriefings reduce belief in false info
⚠️ Both slightly reduce belief in true info
🧠 Only specific boosts perceived learning
🟢 Only general reduces manipulation & frustration
🔍 Both increase transparency
osf.io/preprints/ps...
Reposted by Sacha Altay
More constructive approaches to online political debate generate more constructive replies but no greater change in attitudes www.science.org/doi/pdf/10.1...
July 28, 2025 at 1:46 AM
More constructive approaches to online political debate generate more constructive replies but no greater change in attitudes www.science.org/doi/pdf/10.1...
Reposted by Sacha Altay
How effective are user corrections on social media, and does adding a link to a fact check improve effectiveness?
In piece led by @sachaltay.bsky.social we find corrections have small effects, adding a fact-check unlikely to make them more effective misinforeview.hks.harvard.edu/article/the-... 1/6
In piece led by @sachaltay.bsky.social we find corrections have small effects, adding a fact-check unlikely to make them more effective misinforeview.hks.harvard.edu/article/the-... 1/6
July 23, 2025 at 1:03 PM
How effective are user corrections on social media, and does adding a link to a fact check improve effectiveness?
In piece led by @sachaltay.bsky.social we find corrections have small effects, adding a fact-check unlikely to make them more effective misinforeview.hks.harvard.edu/article/the-... 1/6
In piece led by @sachaltay.bsky.social we find corrections have small effects, adding a fact-check unlikely to make them more effective misinforeview.hks.harvard.edu/article/the-... 1/6
Reposted by Sacha Altay
New blog! ‘Consequences of a skewed discourse around generative AI and elections’. Read why @felixsimon.bsky.social, @oii.ox.ac.uk and @sachaltay.bsky.social, University of Zurich, believe claims about the impact of generative AI on elections have been overblown: bit.ly/46YTZg7 1/2
July 23, 2025 at 10:06 AM
New blog! ‘Consequences of a skewed discourse around generative AI and elections’. Read why @felixsimon.bsky.social, @oii.ox.ac.uk and @sachaltay.bsky.social, University of Zurich, believe claims about the impact of generative AI on elections have been overblown: bit.ly/46YTZg7 1/2
Reposted by Sacha Altay
In addition to the original UK results, we have now ***replicated*** this (TWICE) in the US.
The main findings hold strong: information diets are a lot more diverse in attention than in engagement.
New version here: osf.io/preprints/os...
The main findings hold strong: information diets are a lot more diverse in attention than in engagement.
New version here: osf.io/preprints/os...
July 22, 2025 at 8:27 AM
In addition to the original UK results, we have now ***replicated*** this (TWICE) in the US.
The main findings hold strong: information diets are a lot more diverse in attention than in engagement.
New version here: osf.io/preprints/os...
The main findings hold strong: information diets are a lot more diverse in attention than in engagement.
New version here: osf.io/preprints/os...
Reposted by Sacha Altay
Today (w/ @ox.ac.uk @stanford @MIT @LSE) we’re sharing the results of the largest AI persuasion experiments to date: 76k participants, 19 LLMs, 707 political issues.
We examine “levers” of AI persuasion: model scale, post-training, prompting, personalization, & more!
🧵:
We examine “levers” of AI persuasion: model scale, post-training, prompting, personalization, & more!
🧵:
July 21, 2025 at 4:20 PM
Today (w/ @ox.ac.uk @stanford @MIT @LSE) we’re sharing the results of the largest AI persuasion experiments to date: 76k participants, 19 LLMs, 707 political issues.
We examine “levers” of AI persuasion: model scale, post-training, prompting, personalization, & more!
🧵:
We examine “levers” of AI persuasion: model scale, post-training, prompting, personalization, & more!
🧵: