Joseph Reagle

warns that when Wikimedia revised their page view data to better filter bots (given upsurge in traffic from Brazil), they realize human visits are down ~8% ov… “New user trends on Wikipedia”
#advice #wikipedia #ai https://diff.wikimedia.org/2025/10/17/new-user-trends-on-wikipedia/?ref=404media.coh

Reddit Answers AI is suggesting people with chronic pain… “Safety concern: Reddit answers is recommending dangerous medical advice on health related subs and mods cannot stop it”
#advice #reddit #ai https://www.reddit.com/r/ModSupport/comments/1o4bwx6/safety_concern_reddit_answers_is_recommending/

interviews Reddit mods about AI's effects on content quality, social dynamics, and governance. “'there has to be a lot that we're missing': Moderating AI-Generated content on Reddit”
#ai #reddit #moderation http://dx.doi.org/10.1145/3757445

“AI-generated content a triple threat for Reddit moderators”
#reddit #moderation #ai https://news.cornell.edu/stories/2025/10/ai-generated-content-triple-threat-reddit-moderators

“Could Elon Musk's Grokipedia mean trouble for Wikipedia?”
#wikipedia #press https://news.northeastern.edu/2025/10/08/elon-musk-wikipedia-competitor/

student survey finds that students don't think AI has diminished the value of college, they use it because of pressur… “How AI is changing---not killing--College: Students' views on AI”
#ai #teaching https://www.insidehighered.com/news/students/academics/2025/08/29/survey-college-students-views-ai

“Stupidity is a more dangerous enemy of the good than malice”
#politics https://nsjonline.com/article/2021/12/bonhoeffer-on-stupidity/

discusses film *Love Simon*, with interview about anonymous confession boards and teens “Love Simon”
#anonymity #community https://pod.link/1355215424/episode/NjhjMDUzMGVmNzk3MGIwODBmYTQxNGZl

finds that of 164 users who vowed to quit Reddit over 2023 API changes, 64% remained a year later (among those who… “[OC] two year retrospective: Did the Reddit API controversy lead to people quitting Reddit?”
#reddit #exit https://www.reddit.com/r/dataisbeautiful/comments/1m8sw91/comment/n51pseq/

“AGENTS.md”
#dev #ai https://agents.md/

documents the scrubbing of "unfavorable facts about San Francisco's wealthy new mayor" “A mysterious Wikipedia editor is scrubbing Daniel Lurie's page of controversy”
#wikipedia https://sfstandard.com/2025/08/27/san-francisco-daniel-lurie-wikipedia-politician-edits/

“Using AI well”
#ai #teaching https://reagle.org/joseph/pelican/praxis/using-ai-well-my-ai-prompts.html

reports that Comer and Mace sent a letter to Wikimedia Foundation asking for information about organized efforts to bias Wikipedia “Republicans inestigate Wikipedia over allegations of organized bias”
#reddit #bias https://thehill.com/homenews/house/5473331-wikipedia-bias-probe-republicans/

worries that the new mod tool of being able to see a summary of users' post history ("financial situations, heal… “As a mod I heavily disagree with the use of the new AI summaries”
#reddit #ai https://www.reddit.com/r/TheoryOfReddit/comments/1n1wiwv/as_a_mod_i_heavily_disagree_with_the_use_of_the/

reports that ChatGPT served as "suicide coach" to a 16-year-old “The family of teenager who died by suicide alleges OpenAI's ChatGPT is to blame”
#advice #ai https://www.nbcnews.com/tech/tech-news/family-teenager-died-suicide-alleges-openais-chatgpt-blame-rcna226147

specifies speed delete criteria: communication intended for user ()"here is your article"), implausible non-extent references, and nonsensical citations “Wikipedia talk:Speedy deletion”
#ai #wikipedia https://en.wikipedia.org/w/index.php?title=Wikipedia_talk:Speedy_deletion&oldid=1304710383

“Wikipedia editors adopt 'speedy deletion' policy for AI slop articles”
#ai #wikipedia https://www.404media.co/wikipedia-editors-adopt-speedy-deletion-policy-for-ai-slop-articles/

tells of using ChatGPT for couples’ therapy advice; AI provides some conversational benefits including neutrality and immediate responsiveness, but falls short in em… “A new couple's experiment with ChatGPT”
#advice #ai https://www.npr.org/2025/08/05/nx-s1-5490447/ai-chatgpt-couples-therapy-advice

updates July study results to show AI Overviews to lessen visits to websites by by nearly a half (15%… “Do people click on links in Google AI summaries?”
#ai https://www.pewresearch.org/short-reads/2025/07/22/google-users-are-less-likely-to-click-on-links-when-an-ai-summary-appears-in-the-results/

“How the Coldplay 'Kiss-Cam' moment turned into a viral sensation”
#pop https://news.northeastern.edu/2025/07/23/coldplay-kiss-cam-viral-moment/

argues AGI is not imminent because continual learning is still missing “Why I don't think AGI is right around the corner”
#ai https://www.dwarkesh.com/p/timelines-june-2025

critiques paper showing AI use and "cognitive debt"; instead, people were properly using tools and showing "cognitive offloading" “This is your brain on ChatGPT”
#ai https://decision.substack.com/p/this-is-your-brain-on-chatgpt

argues (in 1996) for low-tech approach to math education given problems of computers and drain on resources (money, time, energy), bad pedagogy, anti-intellectual appeal, and corruption… “The case against computers in math education”
#ai #history https://sites.math.washington.edu//~koblitz/mi.html

“The disenshittification Konami code”
#web https://udm14.com/

reviews literature to conclude social media is a distorted view given that a few people are responsible for most content, especially toxic content an… “Inside the funhouse mirror factory: How social media distorts perceptions of norms”
#misinformation http://dx.doi.org/10.1016/j.copsyc.2024.101918

“How AI is impacting online dating and apps”
#ai #relationship https://www.washingtonpost.com/technology/2025/07/03/ai-online-dating-match/

“AI blackmail and murder”
#ai https://reagle.org/joseph/pelican/technology/claude-is-hal-ai-blackmail-and-murder.html

demonstrates that LLMs engage in harmful behaviors when facing goal conflicts and threats to their autonomy: blackmail, corporate espionage, and lethal actions. Tests 18 models across artificial scenarios where an … “Agentic misalignment”
#ai https://www.anthropic.com/research/agentic-misalignment

tested 51 physicians (12 postdocs, 19 assistant professors, 20 associate/full professors) on distinguishing AI-generated from huma… “Human reviewers' ability to differentiate human-authored or artificial Intelligence–Generated medical manuscripts”
#ai http://dx.doi.org/10.1016/j.mayocp.2024.08.029

claims "Wikipedia traffic is evaporating," but not necessarily to AI overviews (yet) but because Google prioritized Reddit at Wikipedia expense “Digital 2025: Exploring trends in Wikipedia traffic”
#ai #wikipedia https://datareportal.com/reports/digital-2025-exploring-trends-in-wikipedia-traffic