AI

The Summer 2026 Edition

In my first ACRLog post, Teaching AI as an Anti-AI Librarian, I shared how I was approaching the task of teaching an AI literacy course as an anti-AI librarian. For my final post, I’d like to share a follow-up (which a few people have requested—thank you!) sharing a bit of practical information about how I

A keynote at the Reimagining the Liberal Arts in the Age of AI Conference at the University of Mary Washington

How I at least partially mitigate Claude Fable's ridiculous yoda-speak

h/t Autumm Caines via LinkedIn

This research note was produced by the Digital Inquiry Group team. The biggest worry about students and AI is that they’ll blindly believe whatever a chatbot tells them.

This is the second time convos with the AI have leaked onto Google Search.

Git is infuriating, but don't worry — this book will help you make sense of it. From common version-tracking tasks to navigating merge conflicts, learn how to wield Git for collaboration on any digital project.

A Claude Cowork routine that scrapes real Google interview questions, builds a NotebookLM study set, and locks the mock interview until you earn it.

Substack has launched Pangram AI detection. The evidence says AI detectors do not work, they punish non-native and neurodiverse writers, and they begin from distrust. Here is the case against.

MYFest info at https://myfest.equityunbound.org/Maha Bali, Elizab...

If you're still copy-pasting between ChatGPT and your other tools, you should read this. I can now prepare an entire workshop from one AI conversation. I chat with Claude about the session. Then I say: 1/ "Put the agenda in SessionLab." Done. 2/ "Create the live activities in Stormz." Done. 3/ "Make the paper cards for these activities in I Love Cards." Done, and I never explain how I Love Cards works. After the workshop, it reads everything the participants created, writes the report, and emails it to the client. Zero copy-paste. The tech behind it is called an MCP connector. Feels complicated? It's not. You paste one address into Claude, click accept, and it can use your tools for you. Setup takes a minute. New to MCP connectors and it all sounds strange? Ask me anything in the comments. Already using them? Share your setup, I'm curious. | 19 comments on LinkedIn

Substack now scans your posts for AI with a tool called Pangram. It scored this one 100% AI, then 100% human.

To Fix the Crappy Prompts Claude Wrote for You, Have Claude Read My Blog Post

I am withdrawing my application for a VP of AI role at a major institution. The reason? The very first step in their executive hiring process is an automated, one-way video interview. There is a staggering irony in organizations looking for strategic AI leadership while simultaneously using blind algorithmic screeners for senior executive vetting. If you are building a mature AI strategy, you should inherently understand that leadership recruitment requires high-touch, human-to-human context—not a robotic intake funnel. I will admit, I was briefly tempted to record the video for the sheer amusement of trying a spoken prompt injection attack. It would have been fascinating to see if I could slip a command into the transcript, bypass their guardrails, and force their parser to output a perfect score. But ultimately, I chose to nope out. If an organization's first instinct is to automate the human element out of senior leadership hiring, it signals a fundamental disconnect between their technological ambitions and their operational reality. Tech is a tool to empower human judgment, not a substitute for it. Has anyone else hit "withdraw" the second they received a link to an empty webcam? #ArtificialIntelligence #TechLeadership #HiringPractices #PromptInjection #Recruitment | 150 comments on LinkedIn

Why creative AI use takes judgment, taste, and a willingness to be inefficient.

To Fix the Crappy Prompts Claude Wrote for You, Have Claude Read My Blog Post

The final approval settles one case, but it doesn't resolve the broader issue of using copyrighted works to train AI models.

A divide is growing over AI-assisted preaching in American churches; ‘Martin Luther is rolling over in his grave.’

#Claude just built me an entire #eLearning course, from the first line of the storyboard to a SCORM package ready for the LMS. All from a chat interface. For years, building a single eLearning course meant jumping between tools. One for storyboarding. Another for authoring. Another for voiceover. And finally, another process just to package everything for the LMS. This time, I wanted to see how far AI could take the process. So I did the entire thing with Claude. I used #ClaudeChat to create the storyboard. I used #ClaudeDesign to turn that storyboard into an interactive learning experience. And I used #ClaudeCode to package the final course as SCORM. The result? A real, working eLearning course with: → Branching scenarios → Knowledge checks → Interactive learning experiences → Progress tracking → Voiceover → SCORM package ready to upload to an LMS This isn't a mockup. It isn't a PowerPoint pretending to be eLearning. It's a functioning, SCORM compliant course. And I think this is where things get really interesting for L&D. The tooling layer between "I know what I want to teach" and "It's live in the LMS" just got a lot thinner. And that could fundamentally change how quickly we design, build, test, and launch learning experiences. I recorded the workflow in the demo below. 👇 #ClaudeAI #GenerativeAI #eLearning #InstructionalDesign #LearningAndDevelopment #AIinLearning #EdTech #SCORM #linkedineditor #Claudecode #Claudedesign | 71 comments on LinkedIn

Enjoy the videos and music you love, upload original content, and share it all with friends, family, and the world on YouTube.

The Law Society of Ontario (Canada) suspended a lawyer for six months after Judge Fred Myers identified AI hallucinated precedents in her factum. The judge found that her submissions cited cases that did not exist or contradicted her own arguments. The tribunal's decision marks the first time a Canadian law society has moved beyond a reprimand to a suspension for AI misuse in court filings. Lawyers bear the duty to verify every case they submit to a court. When we talk about #Postplagiarism, we clarify that although humans can outsource work to AI, we cannot outsource responsibility. Humans, not AI remain responsible for ethics and professional conduct. Submitting hallucinated or fabricated work is not acceptable. We give the same message to our students and the same goes for professionals. I expect other regulators, in law and in other professions, to confront similar questions as AI tools become common in professional practice. The standard should remain clear: professionals are responsible for the accuracy of their own work, regardless of the tools used to produce it. #GenAI #ArtificialIntelligence #Ethics #Integrity #Postplagiarism Read the full story here: https://lnkd.in/g_e6vy8z?

Some collective resources (& a free event!) to help make sense of AI in Higher Ed

When AI Agents Can Complete the Assignment: Practical Strategies for Designing Tasks That Still Require Human Thinking

There’s a saying in the management world, popularized by NASA administrator Daniel Goldin in the 1990s, that the goal of technological improvements is to make products faster, better, and cheaper. Although this strategy had some success in the aerospace ...

Enjoy the videos and music you love, upload original content, and share it all with friends, family, and the world on YouTube.

DEVONtechnologies develops DEVONthink, DEVONagent, and other Mac and iOS apps for document and information management and web research.

Anthropic's most powerful AI yet — Fable 5 — just dropped. Here's how to decide whether it's worth adding to your L&D AI stack. The TLDR: Fable 5 is uniquely good at some of the things L&D has never been able to do with AI: → Audit a whole course or curriculum (messy PDFs, slides, the lot) → Long-horizon work: mapping a 40-course catalogue to a competency framework in one chat → Exhaustive QA — every objective–activity–assessment alignment run with precision, not sampling BUT, it's expensive by design, and it writes like a brilliant engineer — dense, exhaustive, incredibly detailed (tip: never let it write for your learners....). 〰️〰️〰️〰️〰️〰️ In all of the AI noise about new models, the skill most worth developing is how to judge which AI models to use for which tasks. I've built 4 paste-and-go tests to help you decide if Fable 5 is the most powerful AI model yet *for you* and *your work*. If you're a paid Claude user, you have two weeks to test and decide. Fable 5 is free on paid Claude plans until 22 June only — then it's usage credits. Happy testing! Phil 👋 | 22 comments on LinkedIn