Ezequiel Lopez-Lopez

In light of the assault on American science by its own government that launched in January this year, we published the Anti-Autocracy Handbook in June to provide some guidance to scholars about how to respond to an emerging autocracy. The handbook is available at: sks.to/autocracy. 1/n


1/4

#AI #DigitalHealth #Misinformation #ConfirmationBias #ChatGPT

As GenAI becomes part of daily life, its influence on behavior and decision-making, especially in health, is critical.
🧠 Combating confirmation bias in this space requires:
• Equipping users with critical AI literacy
• Establishing strong oversight, regulation, and auditing

We also call for systemic solutions:
🔍 Public black-box testing
🧑‍⚕️ Reporting tools for physicians to flag GenAI-influenced patient behavior
📋 Regulatory oversight for general-purpose GenAI used in medical contexts

So what can we do? We propose:
✅ Boosting digital/AI literacy
✅ Training people to test multiple query framings
✅ Promoting “consider-the-opposite” thinking
✅ Designing GenAI apps to flag evidence-based but disconfirming info

🧠 GenAI doesn’t just answer; it engages in dialogue. This allows people to iteratively steer the conversation toward preferred narratives, reinforcing cognitive biases without necessarily realizing it

we showcase how small changes in system configuration (like creating a more empathetic chatbot) made GenAI more likely to mirror user bias, endorsing pseudoscientific treatments and downplaying evidence-based medicine

In traditional search (e.g., Google), these biases already existed. But GenAI takes it further through hypercustomization (journals.sagepub.com/doi/10.1177/...), adapting to your wording, preferences, even emotional tone. This can lead to tailored responses that validate your views, accurate or not

We identify 3 key “pressure points” where GenAI can entangle with confirmation bias in health contexts:
1️⃣ How users phrase queries
2️⃣ Preference for belief-consistent answers
3️⃣ Resistance to belief-inconsistent info

Millions are turning to AI tools for medical queries.
While GenAI offers easy access to personalized health information, that personalization can backfire, subtly reinforcing people’s pre-existing beliefs, even when they’re false or harmful

with our wonderful team @cabels18.bsky.social @stefanherzog.bsky.social @lewan.bsky.social @dlholf.bsky.social
@mpib-berlin.bsky.social @arc-mpib.bsky.social @unipotsdam.bsky.social @bristoluni.bsky.social

NYAS Publications Generative artificial intelligence (GenAI) applications, such as ChatGPT, are transforming how individuals access health information, offering conversational and highly personalized interactions. Whi...

New research out!🚨

In our new paper, we discuss how generative AI (GenAI) tools like ChatGPT can mediate confirmation bias in health information seeking.
As people turn to these tools for health-related queries, new risks emerge.
🧵👇
nyaspubs.onlinelibrary.wiley.com/doi/10.1111/...

Opinion | José Andrés: People of Good Conscience Must Stop the Starvation in Gaza

From @nytopinion.nytimes.com

“There is no excuse for the world to stand by and watch two million human beings suffer on the brink of full-blown famine,” the chef José Andrés writes about Gaza.

ChatGPT Is Changing the Words We Use in Conversation Words frequently used by ChatGPT, including “delve” and “meticulous,” are getting more common in spoken language, according to an analysis of more than 700,000 hours of videos and podcasts

Great coverage by @sciam.bsky.social of our research on how ChatGPT is influencing the words we use in conversation!

Work led by Hiromu Yakura & @levinbrinkmann.bsky.social

www.scientificamerican.com/article/chat...

New paper out! 📣
We explore GenAI’s hypercustomization capabilities, the behavioral and governance challenges they may introduce in the foreseeable future, and potential strategies to address them.

You sound like ChatGPT AI isn’t just impacting how we write — it’s changing how we speak and interact with others. And there’s only more to come.

You sound like ChatGPT -- that is what The Verge said about our recent preprint on how ChatGPT is reshaping human spoken communication.

Chatbots are becoming a cultural medium and agent of cultural homogenisation. Is this inevitable, or can model diversity help? www.theverge.com/openai/68674...

@arc-mpib.bsky.social @mpib-berlin.bsky.social @lorenzspreen.bsky.social @anaskozyreva.bsky.social

I am privileged to announce the publication of the Anti-Autocracy Handbook: sks.to/autocracy 1/12

Bust of Andrey Sakharov

Science is under a wide-ranging attack in the U.S. From arbitrary and catastrophic funding cuts to censorship based on keywords, scholarship is no longer free from government interference. How can scientists respond? 1/n

Free speech, fact checking, and the right to accurate information True to his campaign promises, on 20 January 2025, US President Donald Trump signed a broad range of Executive Orders, the scope of which ranged from renaming the Gulf of Mexico to “Gulf of America” t...

My latest column just appeared in Science, entitled "Free speech, fact-checking, and the right to accurate information”. (www.science.org/doi/10.1126/...) I use one of President Trump’s first executive orders to unpack the terrain between misinformation and claims to free speech 1/n

🚨Press release🚨

How can LLMs help and hurt collective intelligence?

Interdisciplinary team of 28 scientists led by our associate researcher @jasonburton.bsky.social proposes recommendations for action (Nature Human Behaviour)

www.mpib-berlin.mpg.de/press-releas...

Identifying the underlying psychological constructs from self-expressed anti-vaccination argumentation - Humanities and Social Sciences Communications Humanities and Social Sciences Communications - Identifying the underlying psychological constructs from self-expressed anti-vaccination argumentation

🚨New paper by Dawn Holford et al., in collaboration with ARC's Ezequiel Lopez-Lopez @eloplop.bsky.social 🚨

This study reveals the psychological "attitude roots" behind negative vaccine views by analyzing 556 self-expressed reasons

👇Check it out for more details! www.nature.com/articles/s41...

AI is even changing how we speak...?

A recent study in collaboration with ARC's Ezequiel Lopez-Lopez @eloplop.bsky.social shows a significant rise in words associated with ChatGPT using 280,000+ YouTube videos.

👇Check out the preprint and explore the feedback loops between AI and human culture.

Empirical evidence of Large Language Model's influence on... Artificial Intelligence (AI) agents now interact with billions of humans in natural language, thanks to advances in Large Language Models (LLMs) like ChatGPT. This raises the question of whether...

🚨 preprint 🚨

Can ChatGPT influence the way we speak to each other? I.e. can they shape human culture?

We transcribed and analyzed 300k YouTube videos to see if humans increased the use of words like 'delve' and 'adept,' which are known to be overused by ChatGPT.

Paper: arxiv.org/abs/2409.01754

Misinformation poses a bigger threat to democracy than you might think In today’s polarized political climate, researchers who combat mistruths have come under attack and been labelled as unelected arbiters of truth. But the fight against misinformation is valid, warrant...

Misinformation remains a threat to democracy.

nature.com/articles/d41...

@lewan.bsky.social @naomioreskes.bsky.social @profsanderlinden.bsky.social

Our first day of SciBeh 's Virtual Workshop 2024 on "Epistemic Boundaries" is almost over!
Missed day 1? No worries! You can still join us for Day 2's insightful talks and discussions. Details and sign-up here:
www.scibeh.org/events/works...

We’ll delve into how expertise can inform policymaking during crises, identify the conditions that limit the applicability of expertise, and examine scenarios where expertise in one area may overreach into other disciplines–a situation we call ‘epistemic trespassing’.

Interested in the boundaries of expertise? Join us at SciBeh on March 6 & 7, 2024 for an engaging virtual workshop.
www.scibeh.org/events/works...