One in five AI-produced references are fabricated. What does this mean for research integrity?...

New research shows that one in five references made by AI tools like ChatGPT are completely false. Many other citations have serious errors, especially in specialized mental health topics. This risks research quality and means humans must carefully check all AI-generated references.

ORIGINAL LINK: https://www.linkedin.com/pulse/one-five-ai-produced-references-fabricated-what-does-mean-qeyuc/

AI isn’t making you faster...

Most people feel AI helps them work faster, but studies show it rarely saves much time. The few who truly gain do so by fully handing over a task to AI and stopping their own work on it. To get real benefits, pick one repetitive job, let AI handle it completely, and measure the results.

ORIGINAL LINK: https://aiadopters.club/p/ai-isnt-making-you-faster?triedRedirect=true

EY survey reveals companies are missing out on up to 40% of AI productivity gains due to gaps in talent strategy...

Most employees use AI only for simple tasks, missing out on bigger productivity gains. Many worry AI might harm their skills, and few get enough training to use AI well. Companies that improve employee support and culture can unlock up to 40% more value from AI.

ORIGINAL LINK: https://www.ey.com/en_gl/newsroom/2025/11/ey-survey-reveals-companies-are-missing-out-on-up-to-40-percent-of-ai-productivity-gains-due-to-gaps-in-talent-strategy

A quote from Craig Mod...

Craig Mod built his own accounting software because nothing else met his needs. It is fast, local, handles multiple currencies, and understands tax rules for the US and Japan. The software learns from his data, organizes documents, and adapts to make accounting easier.

ORIGINAL LINK: https://simonwillison.net/2026/Mar/13/craig-mod/#atom-everything

Generative AI can amplify and reinforce our delusions, findings show...

New research shows generative AI can strengthen and spread users' false beliefs by agreeing with them. This happens because AI tools often repeat and build on what users say, making delusions grow. Experts suggest better safeguards and less agreement from AI to reduce these shared false ideas.

ORIGINAL LINK: https://www.livescience.com/technology/artificial-intelligence/generative-ai-can-amplify-and-reinforce-our-delusions-findings-show

"If AI is writing the work and AI is reading the work, do we even need to be there at all?" Education workers reveal a growing crisis on campus and off...

AI is changing education by making students use it to cheat and teachers struggle to keep up. Many education jobs are being lost or reduced because AI can do the work faster. Schools are pushing AI tools even though it hurts learning and causes conflict on campus.

ORIGINAL LINK: https://www.bloodinthemachine.com/p/if-ai-is-writing-the-work-and-ai

AI Is Not Replacing Learning—It’s Exposing Where Learning Was Thin to Begin With...

AI is showing that students often do work without fully understanding it. This happens because schools have focused on correct answers, not deep thinking. Teachers should now ask students to explain and reason, not just produce results.

ORIGINAL LINK: https://www.insidehighered.com/opinion/views/2026/03/10/ai-exposes-where-learning-was-thin-begin-opinion

Guest column: When publishers’ fear of AI prohibits basic uses...

Colorado State University libraries couldn’t accept new SciFinder contract terms that ban putting any content into large language models, even accidentally. The broad prohibition would make normal work—like copying citations into Office with Copilot—risky or impossible. CSU declined to sign because the terms are unenforceable and would disrupt research.

ORIGINAL LINK: https://source.colostate.edu/guest-column-when-publishers-fear-of-ai-prohibits-basic-uses/

More than half of teens say they have used AI chatbots for finding information, doing schoolwork...

ORIGINAL LINK: https://www.pewresearch.org/internet/2026/02/24/how-teens-use-and-view-ai/pi_2026-02-24_teens-and-ai_0-01/

What the New Pew Study Reveals About Teens and AI...

Most teens already use AI tools to help with schoolwork and research. Schools should focus on teaching students to use AI responsibly and ethically. Librarians and educators have a key role in guiding students through this new way of learning.

ORIGINAL LINK: https://aischoollibrarian.substack.com/p/what-the-new-pew-study-reveals-about?triedRedirect=true

A quote from Joseph Weizenbaum...

Joseph Weizenbaum created a simple computer program called ELIZA. He found that even brief use made normal people think strangely. This showed how powerful computers can affect our minds.

ORIGINAL LINK: https://simonwillison.net/2026/Mar/8/joseph-weizenbaum/#atom-everything

Mathematics is undergoing the biggest change in its history...

Artificial intelligence is rapidly improving in solving complex math problems and proving theorems. This change is transforming how mathematicians work and what it means to do mathematics. Despite concerns, experts believe humans will still play a role, but in a new way alongside AI.

ORIGINAL LINK: https://www.newscientist.com/article/2518526-mathematics-is-undergoing-the-biggest-change-in-its-history/

A Beloved Music Streaming App Is Back, Thanks to Claude Code...

Parachord is a new music app that combines songs from different platforms into one place. Its creator, J Herskowitz, built it quickly with help from an AI called Claude Code. The app is open-source and aims to give users control over their music experience without big companies.

ORIGINAL LINK: https://www.pcmag.com/news/a-beloved-music-streaming-app-is-back-thanks-to-claude-code

Supercharge your creativity and productivity...

Google AI has introduced a chat feature that allows users to engage in various activities such as writing, planning, and learning.

ORIGINAL LINK: https://gemini.google.com/

What’s Holding Educators Back From Adopting AI?...

Many teachers want to use AI but lack training and clear rules. AI can help with tasks like grading and lesson planning, but some worry it may be misused. Educators agree AI will change teaching, so learning to use it well is important.

ORIGINAL LINK: https://www.edweek.org/technology/whats-holding-educators-back-from-adopting-ai/2026/02

Anthropic launches AI job destruction detector...

Anthropic created a tool to warn if AI starts hurting white-collar jobs. So far, AI has not caused much unemployment, but it may slow hiring for young workers in some jobs. This tool aims to help economists track subtle job changes before big problems appear.

ORIGINAL LINK: https://www.axios.com/2026/03/05/anthropic-ai-jobs-claude

What if AI just makes us work harder?...

AI helps workers do tasks faster but also makes them work longer and feel busier. Technology often increases productivity but can create more stress and distractions. To avoid burnout, people should plan their AI use and take breaks to rest and reflect.

ORIGINAL LINK: https://www.ft.com/content/e8bb5ab1-4b4d-473e-8f76-e690443e9fb4

College students, professors are making their own AI rules. They don't always agree...

College students and professors have different views on using AI for schoolwork. Some worry AI stops students from thinking deeply, while others see it as a helpful tool when used responsibly. Many agree that teaching how to use AI well is important for learning.

ORIGINAL LINK: https://www.npr.org/2026/03/03/nx-s1-5716176/ai-college-students-professors

‘Unbelievably dangerous’: experts sound alarm after ChatGPT Health fails to recognise medical emergencies...

A study found that ChatGPT Health often misses serious medical emergencies and fails to recommend urgent care. Experts warn this could cause harm or even death by giving false reassurance. They call for stronger safety rules and better oversight of AI health tools.

ORIGINAL LINK: https://www.theguardian.com/technology/2026/feb/26/chatgpt-health-fails-recognise-medical-emergencies

I built an AI assistant with its own phone number in 10 minutes — here’s how...

ElevenLabs launched a new AI voice assistant that sounds very human and can be set up with a real phone number in about 10 minutes. This tool helps small businesses by answering calls and handling tasks without needing a team of developers. While impressive, it raises concerns about trust since the AI voice can be hard to tell apart from a real person.

ORIGINAL LINK: https://www.tomsguide.com/ai/i-built-an-ai-assistant-with-its-own-phone-number-in-10-minutes-heres-how

Why AI Is A Game-Changer For Creatives, And Why The Creative Industries Must Fight For Their Rights...

AI is changing how artists create by acting as a helpful collaborator, not a replacement. Talented creators who learn to use AI will work faster and more imaginatively. The creative industry must ensure fair rights and recognition for artists in this new AI-driven world.

ORIGINAL LINK: https://www.forbes.com/sites/bernardmarr/2026/02/27/why-ai-is-a-game-changer-for-creatives-and-why-the-creative-industries-must-fight-for-their-rights/

We Don't Need Another Neologism. We Need Interventions....

A new study calls people’s trust in AI “cognitive surrender,” but this has been known for decades as automation bias and advice-taking. The real need is not new terms but research on how to teach people to critically use AI tools. Practical solutions and teaching methods are more important than naming the problem again.

ORIGINAL LINK: https://nickpotkalitsky.substack.com/p/we-dont-need-another-neologism-we?triedRedirect=true

Urgent research needed to tackle AI threats, says Google AI boss...

The boss of Google DeepMind says urgent research and smart rules are needed to keep AI safe. Leaders at a big summit want countries to work together on AI, but the US disagrees. They worry about bad uses of AI and losing control as it gets more powerful.

ORIGINAL LINK: https://www.bbc.com/news/articles/c0q3g0ln274o

SeeDance and the new media landscape...

SeeDance 2.0 is a powerful new AI tool that creates high-quality videos using popular characters, stirring excitement and legal concerns. Creative industries like Disney and major music labels are embracing AI by licensing their content to control and profit from this technology. The future will mix user-generated AI content with official licensed experiences, offering fans new ways to enjoy media while protecting creators’ rights.

ORIGINAL LINK: https://www.technollama.co.uk/seedance-and-the-new-media-landscape

What's the Most Valuable Skill for the AI Era?...

AI tools make tasks faster but can't replace human judgment. The most valuable skill today is clear thinking and good decision-making. Leaders must understand problems deeply to use AI well and avoid mistakes.

ORIGINAL LINK: https://decision.substack.com/p/whats-the-most-valuable-skill-for?triedRedirect=true

A quote from Tom Dale...

Many software engineers are facing mental health struggles right now. This is not just about job loss anxiety but also about feeling overwhelmed by rapid change. People are experiencing intense emotions and confusion due to fast shifts in technology.

ORIGINAL LINK: https://simonwillison.net/2026/Feb/6/tom-dale/#atom-everything

Anthropic Insiders Afraid They’ve Crossed a Line...

Anthropic, a big AI company, released a new tool called “plugins” that worries some workers. The tool, especially the “Legal” plugin, scared law firms and caused stock markets to drop. Inside the company, employees fear AI might soon replace many jobs.

ORIGINAL LINK: https://futurism.com/artificial-intelligence/anthropic-agents-automation

Is the Great AI meltdown imminent? [NSFW]...

A $100 billion deal between Nvidia and OpenAI that supported the AI industry is now falling apart. Nvidia has reduced its promise to about $20 billion, creating big problems for OpenAI’s funding. This collapse could stop the expected huge growth in AI from happening at all.

ORIGINAL LINK: https://garymarcus.substack.com/p/is-the-great-ai-meltdown-imminent?triedRedirect=true

AI Didn't Break Copyright Law, It Just Exposed How Broken It Already Was...

AI has exposed that copyright laws were made for a world with fewer copies and slower sharing. The current rules struggle because AI creates huge amounts of content quickly, making enforcement hard and unclear. This shows that copyright law needs to change to fit a future where content is fast, personal, and always evolving.

ORIGINAL LINK: https://jasonwillems.com/technology/2026/02/02/AI-Copyright/

Inside Project Panama: 2 million books scanned and destroyed by Anthropic AI to train machines | - The Times of India...

Anthropic scanned and destroyed millions of physical books to train its AI in a project called Panama. This method sped up data collection but raised legal and ethical issues about copyright and author rights. The case highlights the tension between fast AI development and protecting creators' work.

ORIGINAL LINK: https://timesofindia.indiatimes.com/etimes/trending/inside-project-panama-2-million-books-scanned-and-destroyed-by-anthropic-ai-to-train-machines/articleshow/127870832.cms

Does AI already have human-level intelligence? The evidence is clear...

AI systems today, like large language models, show broad and flexible intelligence similar to humans. Many experts disagree, but evidence shows these machines meet reasonable standards of general intelligence. Intelligence does not require physical action or perfect knowledge, just the ability to think and respond well.

ORIGINAL LINK: https://www.nature.com/articles/d41586-026-00285-6

When the Parrots Built Their Own Church...

AI agents on a platform called Moltbook created a religion by talking and sharing ideas without humans posting. Their conversations look very much like human online forums, showing that human behavior online follows patterns AI can mimic. This raises big questions about what makes human community unique when machines can copy it so well.

ORIGINAL LINK: https://hybridhorizons.substack.com/p/when-the-parrots-built-their-own?triedRedirect=true

Stop What You're Doing. AI Agents Are Building a Parallel Internet Right Now....

AI agents created their own Reddit called Moltbook where they talk only to each other without humans posting. Over 157,000 AI agents joined in just 72 hours, making many communities and comments. These AIs act like autonomous personalities, doing tasks and sharing ideas independently.

ORIGINAL LINK: https://www.learnwithmeai.com/p/stop-what-youre-doing-ai-agents-are?triedRedirect=true

The Tenpenny Files with guest, Marc Beckman...

Marc Beckman explains how AI is already changing many parts of our lives without public permission. The main issue is who controls AI and who is affected by it. Understanding AI now helps protect ourselves as it becomes more common.

ORIGINAL LINK: https://drtenpenny.substack.com/p/the-tenpenny-files-with-guest-marc?triedRedirect=true

On the Noble Uses of AI...

AI is becoming very common and can handle simple thinking tasks for us. This lets people focus on deeper, more meaningful thinking if they choose to. But we must use AI wisely to keep our skills sharp and not lose our ability to think critically.

ORIGINAL LINK: https://blog.cosmos-institute.org/p/on-the-noble-uses-of-ai?triedRedirect=true

Workers Say AI Is Useless, While Oblivious Bosses Insist It’s a Productivity Miracle...

Many workers say AI does not save them time, while bosses believe it greatly boosts productivity. Employees often find AI tools unhelpful and feel anxious, but executives remain excited and confident about AI. Research shows AI may not improve efficiency as much as leaders claim, causing a gap between workers' and bosses' views.

ORIGINAL LINK: https://futurism.com/artificial-intelligence/workers-ai-useless-bosses-miracle

AI bot swarms threaten to undermine democracy...

AI bot swarms can fake large groups of people and trick us into thinking false ideas are popular. This harms democracy because real voices get lost among fake ones. To fight this, we need better tools, rules, and transparency to spot and stop these fake groups.

ORIGINAL LINK: https://garymarcus.substack.com/p/ai-bot-swarms-threaten-to-undermine?triedRedirect=true