@jevinwest

Calling Bullshit: The Art of Skepticism in a Data-Driven World @CT_Bergstrom and I wrote a general-audience book on how to spot and refute misinformation, especially the kind wrapped in data. Coming Aug. 4th, available here: penguinrandomhouse.com/books/563882/c…

Honored to be quoted alongside my advisor @jevinwest in the latest @Nature news article by @dcastelvecchi! It’s exciting to contribute to the conversation on @SakanaAILabs’s AI scientist and automating science in general. Check it out here: nature.com/articles/d4158… A full commentary is coming soon. Sharing one of the preliminary results related to popularity bias in cited references with AI generated papers disproportionately citing highly cited papers (~26000 citations per paper on average) in comparison to human written papers (~7500 citations) — an observation that was recently made by @jimmykoppel as well. #AcademicTwitter #Science @SciencePlusAI #ScienceOfAIMediatedScience

🧵@jevinwest and I recently wrote a commentary: "Search engines post-#ChatGPT: How generative artificial intelligence could make search less reliable." We discuss 6 critical issues as search engines transition from "helping" you find the answer to "answering" your question.

TrueMedia.org: new non-partisan, non-profit effort to detect manipulated media geekwire.com/2024/new-ai-no… via @GeekWire

New nonpartisan AI nonprofit @truemediadotorg, led by Oren @Etzioni and backed by Uber co-founder Garrett Camp (@gc), is making a political deepfake detector geekwire.com/2024/new-ai-no… via @GeekWire

The Misguided War on the SAT nytimes.com/2024/01/07/bri…

Ironic that Ioannidis led a study on extremely productive authors, which "has become worryingly common across scientific fields." From my count he has published more than 150 papers in the last 2 years. That is about a paper every 5 days. biorxiv.org/content/10.110…

Silicon Valley hype oozes from this video (eg, ‘understanding of science’, ‘reasoning capabilities’, ‘reads 200k papers’) but the re-graphing tool is pretty nice. If it can do that consistently and reliably, worth trying out. youtu.be/sPiOP_CB54A?si… via @YouTube

Press Forward Will Award More Than $500 Million to Revitalize Local News macfound.org/press/press-re…

Frightening how easy disinformation has become: youtu.be/cwGdkrc9i2Y?si… via @YouTube

Have large language models developed analogical reasoning skills? Maybe, but our simple experiments demonstrate otherwise: arxiv.org/abs/2308.16118

Congress escalates push for disinformation researchers’ private documents washingtonpost.com/technology/202…

Suspensions of prominent accounts minimally impact platform engagement osf.io/preprints/soca… via @OSFramework

🧐Can language models generate new scientific ideas? Introducing contextualized literature-based discovery (C-LBD): 1. (L)LMs are given specific contexts/motivations + retrieved *inspirations* from papers💡 2. Goal is to generate text with novel hypotheses. New preprint: 🧵⬇️

Science published a hopelessly credulous news story entitled "Fake scientific papers are alarmingly common" about one of the worst preprints we've seen in years. This is some next-level bullshit. Over at Mastodon, a full thread: @ct_bergstrom/110357259338364341" target="_blank" rel="nofollow noopener">fediscience.org/@ct_bergstrom/…

Searching the web for science: how small mistakes create big problems nature.com/articles/d4158…

The chatbot era: Better or worse off? My two cents, as it relates to misinformation. seattletimes.com/opinion/the-ch… via @seattletimes

Utah’s social media for kids law could be coming to a state near you vox.com/technology/202… via @voxdotcom

Can Humans Find Common Ground? Sure. Just Start With Sea Slugs. nytimes.com/2022/12/09/us/…

"I’ve wanted to type questions into a search bar and, in seconds, read the consensus from decades of scientific research. Such a tool is already in development and is called Consensus." @DKThomp AI is redefining search. Sign up at consensus.app theatlantic.com/newsletters/ar…