"It’s all hallucination, but we only call it that when we notice it’s wrong. The problem is, large language models are so good at what they do that what they make up looks right most of the time." #AI #hallucination #information #ecologies

“Trey and I are attracted to that like flies to honey,” Mr. Stone said. “Oh, that’s where the taboo is? Over there? OK, then we’re over there.”

Our 2017 @nationalacademies.org report on “Human Genome Editing: Science, Ethics, and Governance" (www.nationalacademies.org/our-work/hum...) was designed to prevent exactly this kind of rogue science. One of the many reasons why we need industry-independent regulatory frameworks for emerging tech …

“Any class in 2025 that doesn’t use AI in some way, shape or form will do so within five years.” Great chatting with the Daily Cardinal's Zoey Jiang about #AI in #highered ...

"Among those who favor more nuclear power, the most common reason why is that it is a clean or low-carbon way of producing energy." Interesting new @pewresearch.org data on #nuclear #energy ... www.pewresearch.org/short-reads/...

“Don't protest and don't trespass." The slogan that kicked off the Boston Tea Party …

Let’s not be naive: We need to be very aware of industry influence in computational social science! Through control of data (and experiment) access the situation is even more problematic than in other industries, which makes the data access rights under DSA article 40 particular important!

This nicely reinforces findings from our recent @pnas.org piece in which we argue “that industry players like Meta make significant investments into long-term research streams … to absolve their platforms of responsibility for adverse effects on society or individuals.” www.pnas.org/doi/10.1073/...

We warned against "political hijacking" of #openscience back in 2021 (academic.oup.com/joc/article/...). As @lewan.bsky.social's diagnosis in @science.org shows, ignoring the scholarship in this space for almost five years creates serious vulnerabilities for science and the societies it serves.

Two-thirds of Democrats and four in five Republicans say that "the higher education system in the U.S. is generally going in the wrong direction." www.pewresearch.org/short-reads/... #highered #scipol

“[M]any students graduate without having benefited from talking very much with their teachers and peers, and they stay stuck in ideological bubbles, unwilling or unable to engage with challenging ideas.”

"[F]ederal judges called the Supreme Court’s emergency orders “mystical,” “overly blunt,” “incredibly demoralizing and troubling” and “a slap in the face to the district courts”."

"At ASU, we currently have about 18,000 international student visas ... For Arizona, this has resulted in more than $160 billion in foreign direct investment ... No American student loses his seat to an international student—ASU admits every qualified student who applies."

#Disasters are becoming more frequent and severe, but research on #DisasterManagement remains limited and siloed. A recent @IssuesInST.bsky.social piece advocates for researchers to work together to develop a unified U.S. #DisasterResearch strategy: buff.ly/KLvNNgz

“Brains, treated with the same urgency, may prove a better longer-term investment.“

Long day, filled with very productive (Board of Trustees) meetings at the @klaus-tschira-stiftung.de and catching the closing ceremony of the @hlforum.bsky.social.

Report: You To Be Fired For Reading This Headline About Charlie Kirk

“There's a 25% chance that the future of #AI will go "really, really badly".”

“Meta proposed … [that] [r]esearchers could also write about their findings more vaguely, avoiding terms like “not compliant” or “illegal.”” Another reason why Meta forcing arrangements to have their employees as co-authors on academic publications is so deeply troubling.

"Acceptance of the scientific consensus was very high in the sample as a whole (95.1%), but also in every sub-sample (e.g. no trust in science: 87.3%) ... [P]eople are motivated to reject specific scientific beliefs, and not science as a whole." journals.sagepub.com/doi/abs/10.1...

"Many researchers and educators lack the critical AI tools and resources to investigate fundamental #AI questions and train students."

“Have you ever pretended to hold more progressive views than you truly endorse to succeed socially or academically? An astounding 88 percent said yes.”

Grateful to the Responsible AI Initiative for allowing me to spend time with them this fall. And looking forward to conversations and collaborations with @utah.edu colleagues about innovating responsibly in this deeply disruptive space.

"As a result, on September 30, 2025 this service and the associated software, the AOL Dialer software and AOL Shield browser, which are optimized for older operating systems and dial-up internet connections, will be discontinued." Holy sh*t! 😜

"[A]lthough universities unleashed an age of massive technical innovation, they failed to innovate their own designs to meet the changing needs of society." Most elite universities are ill-prepared for the corrosive amalgam of #AI, societal #polarization, and their own academic echo chambers ...

“[W]hether it's the way you pay your taxes, whether it's through health-care records, whether it's small things that apply to get a permit through international park or a campsite. All of this stuff is going to be part of the AI fabric,” US goes full China on #AI …

"[O]nly 1 percent of Google searches resulted in the users clicking on the link in the AI summary, which takes them to the page Google is summarizing." #AI won't kill the Enlightenment. Enlightenment will LLM itself to death ...

In the U.S., Democrats' confidence in institutions reaches new lows while Republicans' confidence has increased, mirroring which party holds the presidency. Confidence remains high for institutions like small business, the military, and science. news.gallup.com/poll/692633/...

I don’t understand how Wück didn’t bench the goalie after that first half …

In @pnas.org, we outline current challenges #scicomm research faces, incl. social media having all the power over data and issues that come with that. @dietram.bsky.social Press release: attheu.utah.edu/science-tech... Paper: doi.org/10.1073/pnas... @uofuhumanities.bsky.social

Scientists who show humility and recognize that science "doesn't have all the answers" will be more effective communicators with society, argues @dietram.bsky.social. Read more: https://morgridge.org/story/for-effective-science-communication-just-the-facts-isnt-good-enough/ #SciComm

"[T]he terms of what it is like to be a human are about to change in ways that rival the transformations of the Enlightenment or the Industrial Revolution, only much more quickly." One of the better #AI reads, and related to the "accelerated wickedness" of COVID-19 ... slate.com/technology/2...

Better late than never: The correct link with all the articles. My apologies! www.pnas.org/topic/575

“Scientists do a good job of answering the technical questions they think are relevant, about the risks and the benefits, but these are not the questions communities are asking.” Great chatting with @morgridgeinstitute.bsky.social’s Brian Mattmiller about our @pnas.org Colloquium issue …

“Switching languages frequently in the course of a day (or conversation) may be particularly important.”

"Scientists must compete in an attention economy in which they are just one of many voices, and in which ... algorithms now serve as ... gatekeeper." Is #scicomm becoming its own wicked problem? With huge gratitude to my co-authors, excited to finally see this in "print" ... #openaccess

Excited to see the @pnas.org colloquium issue for the 5th @nationalacademies.org "Reimagining #SciComm" colloquium finally being live. #openaccess www.pnas.org/doi/10.1073/...

“They’re like a salad, but less refreshing, or like a sandwich, but less filling—a worst-of-all-worlds Frankenstein’s monster, an indistinguishable food slurry wrapped in edible cardboard.” I miss Gawker-type writing …

A few months ago, Nature published how-to guide for using ChatGPT to write your peer reviews in 30 minutes. This is, of course, a horrible idea. Here’s my response with @jbakcoleman.bsky.social .

As academics use AI to write their talks and presentations, they not only sound more and more like bots themselves, they also produce training material that will further homogenize the linguistic style of future LLMs. Yay!

"[M]igrating @bsky.app users are signaling political separation from an increasingly conservative X and giving up on the idea of a town square that holds all voices simultaneously." I missed this when it first came out. And no, "Musk started it" is not a constructive counterargument ...

"When scientists warn of the harm this does, they risk being seen as part of a scornful anti-MAGA elite ... Meanwhile, genuine problems with the way science works in America are being neglected." Just finally getting around to reading this issue of @economist.com ...

"[O]n January 6, 2021, the American experiment lost its veneer of infallibility. It shifted from being an aspirational North Star to being just one of many democracies struggling with internal tensions and a changing geopolitical landscape. " Thanks to AAPSS for sharing this ...

How might America’s Supreme Court rule in the biggest cases on its docket? We built an AI model to predict the possible outcomes. Find out how it works: https://econ.trib.al/v0WEoBE

Every fall, I wonder if I should continue to teach the Spiral of Silence and climates of opinion in my intro class. And every year, I am glad I do ... www.pewresearch.org/short-reads/...

New paper alert! Published in Social Cognition, people want experts to be intellectually humble, but see all these experts as falling short of this ideal level of intellectual humility. We discuss implications for person perception and public trust. Full paper here: doi.org/10.1521/soco...

What could possibly go wrong?

“This is why I must step away from my work with two federal institutions … that offer guidance about how the Executive and Legislative branches can be stewards of knowledge and create structure to enable discovery, innovation, and ingenuity.”

"It's frustrating that this even needs to be said." Another great piece by @danwphilosophy.bsky.social. What's particularly ironic is that the "Good vs. Evil" dichotomy promoted by some misinformation researchers is itself so deeply authoritarian and anti-academic ...

“Being around less confident people, for example, could make a dog more nervous, while having an anxious dog can exacerbate an owner’s worries. Similar influences have been found in human-human relationships.”