Julian Berger

Why does a worse candidate win? Or an inferior song dominate?

New article with @alexgelas.bsky.social, @pantelispa.bsky.social & Gaël Le Mens.

We show that often once A becomes even slightly more popular than B, people choose A much more often.

www.science.org/doi/pdf/10.1...

Join us in June at @iast.fr Toulouse for a 2-day workshop on social influence and its large scale implications!

Yessss, this point has been vexing me for some time now. The medical community has specialized risk of bias assessment tools for different study types. Time for the social sciences to step up. Tired of reading garbage in garbage out meta analyses.

📣 Applications for the 23rd Summer Institute on Bounded Rationality are now open!

✨Join us in Berlin @arc-mpib.bsky.social June 08–16, 2026, to explore the topic of “Decision Making in the Age of AI”.

✏️ More details + application form (deadline: March 16): www.mpib-berlin.mpg.de/research/res...

Banner image with screenshot of scientific article from nature Medicine, as well as two panels from the study method and results

⚠️ Despite all the hype, chatbots still make terrible doctors. Out today is the largest user study of language models for medical self-diagnosis. We found that chatbots provide inaccurate and inconsistent answers, and that people are better off using online searches or their own judgment.

🚨New preprint! 🚨
"Boosting metacognition in entangled human-AI interaction to navigate cognitive-behavioral drift"
arxiv.org/pdf/2602.01959

Happy to see this new work coming to the surface alongside @cabels18.bsky.social @lorenzspreen.bsky.social @lewan.bsky.social and @stefanherzog.bsky.social
🧵

Across seven incentivized experiments and a large reanalysis, we systematically manipulated the presence and type of post-choice feedback in repeated risky decisions to test whether feedback shapes behavior through learning mechanisms or through anticipatory changes in preferences.

Across seven incentivized experiments and a large reanalysis, we systematically manipulated the presence and type of post-choice feedback in repeated risky decisions to test whether feedback shapes behavior through learning mechanisms or through anticipatory changes in preferences.

Very happy to see our ice-fishing paper on the cover of @science.org this week! 🎣🎉

We tracked large groups of Finnish competitive ice-fishers to study how social foragers use social information when searching for resources. 🐟

Link: www.science.org/doi/10.1126/... (contact me for open access)

Scientific production in the era of large language models With the production process rapidly evolving, science policy must consider how institutions could evolve

Scientific productivity is increasing for those scientists that adopt LLMs for writing: www.science.org/doi/10.1126/...

❄️🎣 What ice-fishing competitions reveal about human decision-making. New paper in @science.org by researchers from @mpib-berlin.bsky.social, @scioi.bsky.social, and the University of Eastern Finland.

👉 Read more: www.mpib-berlin.mpg.de/newsroom/new...

📄 Paper: www.science.org/doi/10.1126/...

Interviewpodcast "Nur eine Frage": Zerstört das Internet die Demokratie, Ralph Hertwig? Der Psychologe Ralph Hertwig erforscht, wie das Internet Demokratien verändert. Er kennt die Gefahren von Fake News und Hetze – aber auch die Chancen globaler Vernetzung.

Die #Algorithmen vieler #SocialMedia - Plattformen belohnen Aufregung. Desinformation, Hass und Hetze dominieren. Zerstört das Internet die #Demokratie? Darüber spricht Ralph Hertwig @mpib-berlin.bsky.social im Podcast "Nur eine Frage" 🎧 Eine Hörempfehlung -> www.zeit.de/wissen/2026-... @zeit.de

On Evaluating Cognitive Capabilities in Machines (and Other "Alien" Intelligences) (Apologies for the length of this post, which means it gets cut off in the email version.

My latest on Substack -- a write-up of the talk I gave at NeurIPS in December.

aiguide.substack.com/p/on-evaluat...

WARN-D machine learning competition is live » Eiko Fried If you share one single thing of our team in 2026—on social media or per email with your colleagues—please let it be this machine learning competition. It was half a decade of work to get here, especi...

After 5 years of data collection, our WARN-D machine learning competition to forecast depression onset is now LIVE! We hope many of you will participate—we have incredibly rich data.

If you share a single thing of my lab this year, please make it this competition.

eiko-fried.com/warn-d-machi...

Machine+Behavior Conference Welcome to the forefront of behavioral science in the digital age.

Machine+Behavior is back! Pls spread the word.

18-19 May
Submission 8 Feb

Speakers include: Tom Griffiths, Allan Dafoe, Thore Graepel, Laura Weidinger, Kinga Makovi, @baronca.bsky.social, @drkatedevlin.bsky.social, Emilio Calvano, Brian D. Earp, Christopher Summerfield

machinebehavior.science

I am hiring a postdoc for a DFF-funded project on social influence, and the decision processes that fuel rich-get-richer dynamics in the online/offline world. The position is for up to a year, competitive Danish salary, remote work possible. Interested or know somebody? DM me or share!

An aerial photograph of the Tilburg University campus.

I am hiring PhD candidates to study the psychology of attention & technology use at @tilburg-university.bsky.social.

We're looking for motivated & curious scholars with expertise in cognitive psychology and statistics, and offer a friendly work environment with great terms & benefits.

tiu.nu/22989

Come and check out my talk on combining human and AI decision making capabilities ✨

🚨 New preprint - Advancing Cognitive Science with LLMs

Excited to share this new preprint together with Rui Mata. We review how 🤖 LLMs can help address long-standing problems in cognitive science, highlighting opportunities and pitfalls.

🔗 Preprint: arxiv.org/abs/2511.00206

The Cube Rule Identify your food based on the location of structural starch. Is a hot dog a sandwich? No, it's clearly a taco.

Confirmed by the cube rule cuberule.com

The Illusion of Readiness: Stress Testing Large Frontier Models on Multimodal Medical Benchmarks Large frontier models like GPT-5 now achieve top scores on medical benchmarks. But our stress tests tell a different story. Leading systems often guess correctly even when key inputs like images are r...

👀

www.arxiv.org/abs/2509.18234

Would you let AI cheat for you?

Our new paper in @nature.com, 5 years in the making, is out today.

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

Humans Perceive Wrong Narratives from AI Reasoning Texts A new generation of AI models generates step-by-step reasoning text before producing an answer. This text appears to offer a human-readable window into their computation process, and is increasingly r...

When reading AI reasoning text (aka CoT), we (humans) form a narrative about the underlying computation process, which we take as a transparent explanation of model behavior. But what if our narratives are wrong? We measure that and find it usually is.

Now on arXiv: arxiv.org/abs/2508.16599

Können Maschinen Emotionen empfinden, Levin Brinkmann? Levin Brinkmann erforscht am Max-Planck-Institut für Bildungsforschung, ob Maschinen Gefühle zeigen sollten.

Können Maschinen Emotionen zeigen? Im Podcast „Ach, Mensch!“ spricht @levinbrinkmann.bsky.social, was passiert, wenn Maschinen Mitleid äußern, wütend werden oder sich erschrecken und warum Emotionen auch Maschinen helfen könnten 🤖 detektor.fm/wissen/ach-m...

#KI #Podcasts

Our (@malte.the100.ci, @ianhussey.mmmdata.io) commentary "cognitive dissonance in large language models is neither cognitive nor dissonant" has just been accepted for publication at PNAS!

Check out the thread below for a summary of our arguments.

osf.io/preprints/ps...

Fundamentally: the fact that LLMs' output is linguistic has a MASSIVE attention capture over social scientists in particular.

We need to keep our wits about us when we use these models and avoid anthropomorphizing them purely because of formal similarities in language.

osf.io/preprints/ps...

Ohhhh Molly Crockett‘s @mjcrockett.bsky.social keynote talk at #FAccT2025 was soo good 🔥🔥🔥

She talked about how techno-optimism is really human pessimism, how DEAD benchmarks don’t capture full human capacities and feed the hype cycle, how we need to avoid monoculture & imagine new worlds together

🚨What is SOTA on tabular data, really? We are excited to announce 𝗧𝗮𝗯𝗔𝗿𝗲𝗻𝗮, a living benchmark for machine learning on IID tabular data with:

📊 an online leaderboard (submit!)
📑 carefully curated datasets
📈 strong tree-based, deep learning, and foundation models

🧵

A recently published meta-analysis in Nature Human Behaviour "found evidence supporting the efficacy of social comparison as a behaviour change technique in shaping behaviour in the desired direction".

I was curious, so I re-analyzed the manuscript, but the funnel plots below say it all.

🚨Why behavioral public #policy should move from #nudging to #boosting

New podcast episode of @arc-mpib.bsky.social's podcast "Unraveling behavior", hosted by @anasofiamorais.bsky.social talking to Ralph Hertwig

You can find links to where to listen it, show notes & more at: tinyurl.com/3yvfwxc8

(1/4) Ever wondered what tech policy might look like if it were informed by research on collective intelligence and complex systems? 🧠🧑‍💻

Join @jbakcoleman.bsky.social, @lukethorburn.com, and myself in San Diego on Aug 4th for the Collective Intelligence x Tech Policy workshop at @acmci.bsky.social!