Stephen Aguilar

It's first-proof-lookin'-at-Wednesday! A new holiday I just made up. Delighted to report that yes, it contains all my favorite typos. Thank god for a second set of human eyes!

My author copies are here, and I've decided to make them unique by signing and embossing them with a custom stamp. I thought it would be nice for folks to receive something special, especially in a world filled with synthetic products made by AI. Cool poster by @reforest-kelley.bsky.social

AI didn't break your course. It stress-tested it. In today's blog post, @stephenaguilar.com explores how generative AI is challenging traditional assessment in higher education and offers a practical framework for teaching staff who are wondering where to begin. Read on: bit.ly/3TYFm83

Are we asking the right questions when it comes to AI in higher ed? @stephenaguilar.com’s “Authenticating Intelligence” explores how gen-AI challenges traditional assessment and forces educators to rethink meaningful learning. Read the blog: https://bit.ly/4hlYW88 🍎 #AcademicSky #BookSky

The same AI tool can produce brilliant work or useless slop. The difference is rarely the tool. Heck, even us experts get fooled by a confident wrong answer, especially if it's at the edges of what we know. Hopefully soon we'll learn that AI doesn't lower the bar for expertise. It raises it.

LAUSD just named its third superintendent in four years, but that's not even the headline. The last one resigned amid an FBI probe into a $6.2 million AI chatbot deal. The company folded. Half a million students got left holding the bag.

My book Authenticating Intelligence comes out in under two months. My editing hack: read every page out loud. Awkward sentences sound ten times worse spoken. Brutal, but it was worth it. Did it catch every typo? Nope. Probably added a few. Perfection's overrated.

A German court just held Google responsible for false info from AI Overviews. Good. Treat AI like food or cars. Set tolerances. Measure real outputs. Publish the error rates companies bury in fine print. www.wired.com/story/a-cour...

EdTech keeps selling promises. Administrators keep buying without asking hard questions. What do you need to know before investing in AI for your school? I laid it out. www.the74million.org/article/what...

NSF is cutting basic science 20-30% to fund a $1.5B AI initiative. That's cannibalism. Google, CRISPR, barcodes—none funded with a commercial application in mind. Basic science is the foundation of the AI revolution. Gut it now. The breakthroughs stop too. www.nsf.gov/impacts/google

My latest for @the74.bsky.social. As EdTech continues to sell promises, administrators and school leaders should be asking these questions. www.the74million.org/article/what...

Great to hear @mattbernacki.bsky.social talking to @stephenaguilar.com about the power of learning analytics for understanding and supporting student success, as well as how #GenAI can support such work. podcasts.apple.com/us/podcast/a...

Authenticating Intelligence still isn’t out, but it’s good to see it indexed!

Authenticating Intelligence has 100 plays on Spotify! It may not be a lot, but we all have to start somewhere. Plus, I have similar numbers on YouTube and Apple Podcasts. Give it a listen. Spotify: bit.ly/AuthIntelSp Apple: apple.co/4a4lVjG YouTube: bit.ly/AuthIntelYT

Stay tuned for a report from @uscrossier.bsky.social and my center that will help folks navigate these things better.

Oh good. More AI stuff founded on gut feelings because there isn't enough AI stuff already.

I mean self-indulgent love powered by sycophantic AI is still a type of love, I suppose. We just used to think that it was a bad thing.

Join my this Sunday! Creativity and Curiosity in the Age of AI: From the Arts to Education Sunday, 1:30 p.m., Wallis Annen-berg Hall, presented by USC

"There are some baseline tasks that can be offloaded, but we can’t just, you know, undercut all of the entry-level positions, because then there won’t be people who develop the experience to really understand what happens when AI fails.” thehill.com/homenews/edu...

From Authenticating Intelligence:

It’s not impossible to build a distributed community. I’m in many outside of my profession. The trouble is that in person communities are superior when it comes to learning, and it’s proven tough to re-establish them post-pandemic.

Join me Sun, 4/19-1:30pm - 2:30pm at the @latimes.com Festival of Books! Session Title: Creativity and Curiosity in the Age of AI: From the Arts to Education Grab tickets online at latimes.com/fob starting Sunday, April 12, or at ticket booth #390 on the day of, while supplies last.

Given the current climate, Sal can be wrong infinity times—it won't matter. That said, calling out EdTech over hype that is never realized is like shooting fish in a barrel. It's their job to sell things, not make a difference. I just wish more folks knew that.

Listen us academics reach DOZENS of people with our work.

I have similar thoughts on Ed Research. It's easy enough to find disparities/inequities and show that they are statistically significant and/or people talk about them coherently (e.g., Qual work). But such findings are meaningless unless they're paired theories of action or contribute to theory.

This week on the Authenticating Intelligence Podcast I sit down with Dan O’Brien (former Blue Sky Studios animator) to talk about the "homogenization" of animation and the quiet threat to entry-level craft. youtu.be/dIY99L0zPvM

Creativity isn't simply taking that initial spark and making it "real." It's refining; rendering the clumsy into expression. It takes time. Sure, you can try to prompt your way there, but that process is crude—especially if you don't really know what you're doing.

Nothing humans make is neutral. That includes EdTech. Tools are extensions of our cognition, and make plain what we value when we use them to engage with the world around us. I dig into this concept a fair bit in Authenticating Intelligence.

Thank you University of Oxford! 🐝🌎

I felt the same. Also Project Hail Mary is one of the best movies I've seen in a while. It's ALSO very true to the book (won't spoil it). If you liked the book you'll really enjoy this movie. Breath of fresh air.

ICQ was the best.

LinkedIn thoughts about AI are weird. Heading: "Imagine if we used computers the way many organizations use AI today." I guess I need *checks desk* two more screens to fully integrate computers into my workflow. I'm so close!

Sometimes it's good to unplug and paint.

Pre-order is live! a.co/d/00VJtuU2

Here's the final book jacket for Authenticating Intelligence. Super happy with how it came out, and thrilled to have some wonderful endorsements from folks I respect.

Stop the presses! More hours spent trying to learn with support from an educator leads to more learning. 🤯 The fact that there are folks who disagree—or ignore—this fact is shameful. I'm glad folks like Morgan exist to call them out on it.

Nothing says "take me seriously" like being barefoot in a fancy looking bar wearing your best shades of puke green.

AI is moving fast. The US school system? Not so much. I’m joined by @mpolikoff.bsky.social on the Authenticating Intelligence Podcast to discuss how AI is shaking up our locally controlled (and very fragmented) education system. youtu.be/274C1RpXOMY

This sums up my feelings about the pice in Nature pretty well.

May I suggest: Release date = 8/6/26

A very good read for people who don't understand the federal role in research and its importance. // "'I came because America was the place to do science,' he said." www.nytimes.com/2026/03/13/u...

"The new wealth gap is who cares about intellectual integrity and who doesn't. Here are 17 prompts that..."

From chapter 9 of Authenticating Intelligence:

Facts

Anthropic came out with a report. Sectors that keep shoving AI into every nook and cranny predict that AI will *checks notes* be incorporated into a bunch of stuff. I did my own back of the envelope calculations and came up with this. I'm entertaining offers from VCs. Bidding starts at 500 billion

In my interview with EdSoruce I highlight a key argument from Authenticating Intelligence: Educators should understand "AI as a constellation of tools, a series of tools, and then they can decide what problems those tools are going to be used to address.” edsource.org/2026/califor...

One more chance to find typos. Spoiler: I will still miss a couple. I may even add a few, who knows.

Good. This "feature" was dumb.

It's a first for the @uscedpolicyhub.bsky.social Substack - we are sharing brand new, original research results! @annarsaavedra.bsky.social, Amie Rapaport and I are looking into reasons for teen absenteeism using the Understanding America Study edpolicyhub.substack.com/p/the-driver...