Quick one for the educators here.
Somewhere in the last few months the federal government settled on an official definition of AI literacy, and it is already flowing into state bills across more than two dozen statehouses. Most of us never saw it happen.
The definition lands on three things: understand how AI works, evaluate outputs critically, use tools ethically. Reasonable on paper. But it was built on a workforce readiness frame, which optimizes for output. The thing that actually protects a student is not output, it is judgment. The ability to look at a fluent, confident, wrong answer and know it is wrong.
So I am curious about your reality: when a student hands in something polished but hollow, where do you think the gap actually is? Skills, or judgment? And what is one move you have made that built the judgment side?
I wrote up the full breakdown and a one question fix you can add to a lesson this week here: https://www.harvestkernel.com/ai-literacy-definition-what-it-means-for-educators/
New on the blog: The Two-Track Workforce.
The largest staffing firm in the world just released data showing AI hasn't simply destroyed jobs. It's split the workforce into two tracks. One is shrinking. The other is promoting people 3x faster and paying 25 percent more.
The post breaks down what actually gets you on the right track, and why AI fluency alone isn't the answer. Full read linked below. Curious what stands out to you.
https://www.harvestkernel.com/?p=2769
New on the blog: Your AI Policy Is Not AI Literacy
Most schools now have an AI policy, or are about to. Far fewer have a faculty that can confidently use AI to build a rubric, redesign an assignment, or coach a student through responsible use. That gap between the document and the skill is the thing almost nobody is measuring.
Around 85 percent of teachers used AI last school year, but fewer than half got any training. The policy is new. The untrained usage is already old.
Today's post breaks down the Policy-Literacy Gap and how the four SeedStacking phases turn becoming AI literate into a sequence of fifteen-minute wins.
New on the blog: students are booing AI off commencement stages this graduation season. Eric Schmidt got hissed at. A UCF speaker was drowned out. And yet the same students are using AI every single day.
That contradiction is not hypocrisy. It is the clearest signal we have about what AI literacy education keeps failing to deliver. In todays post I name the real problem, the Confidence Gap, and walk through how it actually closes.
Read it here: https://www.harvestkernel.com/students-boo-ai-confidence-gap/
New on the blog this morning.
Norman Eng Just Gave Faculty the Right Questions About AI. Here Is What Comes After Yes.
Norman Eng dropped a piece in Faculty Focus today that every educator should read. Four evidence-based questions for deciding when AI belongs in a lesson. The questions are sharp.
What's missing is what happens once you say yes. That's where SeedStacking fills the gap, and that's what today's post walks through.
Read it. Then tell me which of Eng's four questions is hardest to answer honestly in your own course design.
New on the blog: The Evidence Test: How to Tell AI Education Promises From AI Education Hype
A vendor sends you a deck. The first slide promises personalized feedback, improved performance, confidence. You have a decision to make.
Before any of that, there is a quieter decision first: are the claims real, or are they marketing?
I borrowed a standard from clinical practice. The bigger the claim, the stronger the evidence needs to be. Then I built a 4-question filter you can apply to any AI education promise before you adopt it, recommend it, or build a curriculum around it.
The four questions:
- What exactly is the claim?
- Does the evidence match the size of the claim?
- Does it build a skill or hide one?
- Who is the evidence FROM, and who is the evidence FOR?
The post walks through each, plus a worked example on the "AI Tutor" pitch you have probably seen this month.
Read the full post:
https://www.harvestkernel.com/evidence-test-ai-education-hype/
Second post today, and I think this one is worth your full read:
The Studies Are In: AI Use Does Not Wreck Your Brain. Passive AI Use Does.
Four 2026 studies just converged on the same finding. The variable is not whether you use AI. It is HOW.
The MIT study (54 students, EEG, 4 months) found that 83% of LLM users could not quote a single passage from essays they had just written. Oregon State found a 66% drop in reflection per unit of AI reliance. Baldeo at Middlesex published in the APA journal showing the cognitive scaffolding from trying is exactly what gets lost when you skip the trying. Stanford's research found a single five-minute intervention reliably preserves evaluative thinking: a self-assessment rubric BEFORE the AI prompt.
The reframe in the post:
. AI use is not the variable. The pattern of engagement is.
. Active engagement preserves cognition. Passive offloading accumulates debt.
. AI is excellent for extraneous load. Disastrous for intrinsic load.
Three concrete moves you can use Monday morning are in the post.
Read the full breakdown:
https://www.harvestkernel.com/cognitive-debt-research-active-vs-passive-ai-use/
New on the Harvest Kernel blog:
Europe Just Made the AI Education Decision the US Has Been Avoiding
On May 11, the EU Council approved its first formal conclusions on AI and teaching. All 27 member states agreed on something the US has been avoiding: teachers stay at the center.
Three things in the document that should land for every educator reading this:
-
"Digital humanism" as the values frame. Technology supports human agency. When efficiency competes with autonomy, autonomy wins.
-
Teachers as co-designers of AI tools, not just end users. That changes who has a seat at the table when EdTech vendors build the next product.
-
AI literacy as foundational infrastructure. Not optional. Not nice-to-have. The work the rest of the framework depends on.
Read the full breakdown, including why this matters even if you teach in Milwaukee, not Brussels:
https://www.harvestkernel.com/eu-council-ai-education-human-centered-teachers/
New post: When Students Confess They Used AI, You Are Holding a Teaching Moment
An MIT fiction professor walked into class one Tuesday, asked his students to write by hand with no devices, and what followed changed how he teaches.
One by one, his students started confessing what they had been turning in for months. Most of them had been using AI. Some for ideas. Some for outlines. Some for entire essays.
The instinct in that moment is to read it as a discipline event. The teacher who made the difference read it as something else: the first authentic engagement of the semester. A curriculum opportunity walking through the door.
This week's post unpacks what he did next, why the "gloriously flawed" student sentence still matters, and three concrete moves any teacher can make Monday morning to turn confession into curriculum.
What is your default reaction when a student admits to AI use? Drop a comment below.
The U.S. Department of Education just finalized a supplemental priority that makes AI literacy a competitive factor in federal grant applications. Effective May 13, 2026, this changes the funding landscape for K-12 and higher education.
What it is: a policy lever the Secretary can attach to grant competitions. What it is not: a mandate, a regulation, or a new program. But the signal is unmistakable. Schools that can demonstrate AI literacy integration will have a measurable advantage when competing for discretionary funding.
The tension? 81% of educators say they do not have the time to develop an AI training curriculum, and 75% say they do not have the knowledge. Federal policy just validated the destination. The question is who builds the path.
Read the full analysis: https://www.harvestkernel.com/federal-ai-literacy-funding-priority-department-education/