Found 2709 bookmarks
Newest
Cal State struck a deal with OpenAI. Some students and faculty refuse to use it
Cal State struck a deal with OpenAI. Some students and faculty refuse to use it
When California State University paid OpenAI $17 million last year to give campuses unlimited access to a high-powered educational version of ChatGPT, the goal was to help students learn to use artificial intelligence for their education and future careers. However, the announcement came as a surpris
·flip.it·
Cal State struck a deal with OpenAI. Some students and faculty refuse to use it
Economists publish mathematical proof that AI will destroy the economy
Economists publish mathematical proof that AI will destroy the economy

The model outlines a dangerous feedback loop. Individual companies that replace workers with AI reduce labour costs and gain short-term efficiency. But when this strategy is repeated across an entire economy, displaced workers lose wages and cut spending. Since workers are also customers, aggregate demand begins to shrink.

As consumer spending falls, firms respond rationally by cutting costs even further — usually through more automation. The cycle then accelerates: layoffs reduce demand, falling demand encourages more layoffs, and the process becomes self-reinforcing.

In effect, every firm behaves logically in isolation while collectively driving the economy toward systemic failure.

·flip.it·
Economists publish mathematical proof that AI will destroy the economy
Understanding the Evidence Base on AI in K-12 Education | SCALE Initiative
Understanding the Evidence Base on AI in K-12 Education | SCALE Initiative
AI tools are arriving in schools faster than research can evaluate them. Teachers are experimenting with new tools and districts are writing policies, all while students are already using AI both inside and outside the classroom. But for many education leaders, a basic question remains: What does rigorous research actually say about how AI affects teaching and learning? To help answer that question, we released a new report: The Evidence Base on AI in K-12: A 2026 Review. The report reviews the current research, focusing specifically on studies that convincingly estimate causal impact, meaning studies that can tell us whether an AI tool changed outcomes for students or educators.
·scale.stanford.edu·
Understanding the Evidence Base on AI in K-12 Education | SCALE Initiative
Mayo Clinic AI detects pancreatic cancer up to 3 years before diagnosis in landmark validation study - Mayo Clinic News Network
Mayo Clinic AI detects pancreatic cancer up to 3 years before diagnosis in landmark validation study - Mayo Clinic News Network
Mayo Clinic’s REDMOD AI model can detect hidden signs of pancreatic cancer on routine CT scans up to three years before diagnosis, potentially opening a path to earlier and more curative treatment.
·newsnetwork.mayoclinic.org·
Mayo Clinic AI detects pancreatic cancer up to 3 years before diagnosis in landmark validation study - Mayo Clinic News Network
AI_ClassNotes_#13.m4a
AI_ClassNotes_#13.m4a

AI Class Notes #13: Ground It Before You Generate

Thanks so much for your patience! Please find a revamped AI Class Notes focusing on the kinds of topics you shared in your response to the recent poll. That is, top ranked areas you voted for included 1) AI Productivity; 2) Resource Shares; and 3) Quick Tips. These issues will appear once a week, so check back next Wednesday.

1. AI Productivity

Drop your district's unit rubric (PDF) into NotebookLM. Have it convert the rubric into a student-friendly checklist with kid-language descriptions of each criterion. This can reduce a 30-minute rewrite to under five minutes.

Prompt:

Create a student-friendly checklist from this rubric. Use plain language at a 6th-grade reading level. Turn each criterion into one row with a "Looks like / Sounds like / Feels like" example. Add a final row for student self-assessment.

Source: https://alicekeeler.com/2026/04/11/5-amazing-features-of-notebooklm/

Want access to a 50 image prompt library appropriate for K-12? Check out this free resource compiled for you. Here's an image generated from a rubric prompt in the collection

2. Resource Shares

Matt Miller's free AI Teacher Toolkit, a 25-page PDF with copy-paste prompts, lesson ideas, parent-communication scripts, and student "by the way" lessons. No tool to learn, no account to build. Open the PDF, grab a prompt, paste it into whichever AI assistant your district has approved. Email signup required to download, but the resource itself is free and works in ChatGPT, Gemini, Claude, or Copilot.

Source: https://ditchthattextbook.com/ai-toolkit/

But wait, here's another resource for you:

See the rest online

3. Quick Tips

Open any NotebookLM notebook → Chat panel → Configure Chat. Drop in a one-line teacher persona: who you teach, what unit you're on, what tone you want. Every Studio output after that, including quizzes, study guides, audio overviews, follows that frame instead of giving generic responses. Two minutes of setup, every output sharper from then on.

Sample instruction:

You are helping a 9th-grade biology teacher mid-unit on cell respiration. Match a high school reading level, prefer concrete examples over abstract theory, and end every response with one comprehension-check question.

Source: https://www.jeffsu.org/notebooklm-changed-completely-heres-what-matters-in-2026/

·drive.google.com·
AI_ClassNotes_#13.m4a
Anthropic’s Little Brother
Anthropic’s Little Brother
OpenAI is racing to catch up to its greatest rival.OpenAI does not like to be left out. The week after Anthropic announced Claude Mythos Preview—an AI model that has put governments around the world on edge because of its potential ability to hack into banks, energy grids, and military systems—OpenAI
·flip.it·
Anthropic’s Little Brother
Claude-powered AI coding agent deletes entire company database in 9 seconds — backups zapped, after Cursor tool powered by Anthropic's Claude goes rogue | Tom's Hardware
Claude-powered AI coding agent deletes entire company database in 9 seconds — backups zapped, after Cursor tool powered by Anthropic's Claude goes rogue | Tom's Hardware
PocketOS founder blames ‘Cursor running Anthropic's flagship Claude Opus 4.6’ plus Railway’s infrastructure for data disaster.
·tomshardware.com·
Claude-powered AI coding agent deletes entire company database in 9 seconds — backups zapped, after Cursor tool powered by Anthropic's Claude goes rogue | Tom's Hardware
How to defend yourself against AI cheating accusations
How to defend yourself against AI cheating accusations
Don't panic. Take action.So you've been accused of using artificial intelligence to cheat at school — and you're innocent.You must now somehow prove, despite your instructor's suspicions and the alleged evidence, that you completed the assignment or exam on your own. Yet without convincing proof, suc
·flip.it·
How to defend yourself against AI cheating accusations
[2603.17373] SafeTutors: Benchmarking Pedagogical Safety in AI Tutoring Systems
[2603.17373] SafeTutors: Benchmarking Pedagogical Safety in AI Tutoring Systems
Large language models are rapidly being deployed as AI tutors, yet current evaluation paradigms assess problem-solving accuracy and generic safety in isolation, failing to capture whether a model is simultaneously pedagogically effective and safe across student-tutor interaction. We argue that tutoring safety is fundamentally different from conventional LLM safety: the primary risk is not toxic content but the quiet erosion of learning through answer over-disclosure, misconception reinforcement, and the abdication of scaffolding. To systematically study this failure mode, we introduce SafeTutors, a benchmark that jointly evaluates safety and pedagogy across mathematics, physics, and chemistry. SafeTutors is organized around a theoretically grounded risk taxonomy comprising 11 harm dimensions and 48 sub-risks drawn from learning-science literature. We uncover that all models show broad harm; scale doesn't reliably help; and multi-turn dialogue worsens behavior, with pedagogical failures rising from 17.7% to 77.8%. Harms also vary by subject, so mitigations must be discipline-aware, and single-turn "safe/helpful" results can mask systematic tutor failure over extended interaction.
·arxiv.org·
[2603.17373] SafeTutors: Benchmarking Pedagogical Safety in AI Tutoring Systems