AI’s impacts on learning

David Zetland

“AI’s impacts on learning“

AI is changing nearly everything digital around us, for better (search; analysis) and for worse (cybercrime), but its impacts on education are, I think, much larger than many people understand.

Learning is more about process than output. I wrote this to my students:

Why you should avoid AI¹
Because you, unlike me², are at LUC to learn, and the process of researching, drafting, analyzing and writing is where the learning occurs. Put differently, there are inputs (your time), outputs (your grade) and outcomes (your learning). “Saving time” with AIs may limit your input and give a good output, but it doesn’t deliver the outcome.

To clarify the [¹], I wrote that AIs are good as tutors and search engines. If you instead use them to summarize, compare, analyze, structure, edit or revise your writing, then you are (a) cheating and (b) not learning.

Students encountering a topic for the first time don’t necessarily know what’s right or wrong, which is why I had to subtract way too many points on essays full of AI-slop. I was particularly fierce because

(a) students did not have permission to use AIs and

(b) only a few disclosed AI use. The majority started their confessions with “to be perfectly honest…”  WTF dude! When and where did I ever say LIE TO ME!?

Let’s step back and put some structure on this issue:

1. Students go to school because (a) their parents say so, (b) employers want to see paperwork, and (c) they want to learn. Intrinsic motivations (c) are useful. The extrinsic ones (a & b) are worrying
2. Schools use grades to incentivize extrinsic motivation, but grades (outputs) have always been a poor proxy for outcomes. (They are even worse when it comes to predicting life success, but that’s another post.)
3. Students with intrinsic motivation do not need grades (read this long but powerful illustration [PDF]). Those who lack it look for shortcuts.
4. AI offers a powerful shortcut, writing essays in a few minutes that would take a normal student hours or days of searching, reading, synthesizing, confusion, and mistakes — all for an output that maybe the professor read, but probably nobody else will read.³
5. Students who never liked the idea of grades or cared about the topic have seized on AIs as a way of taking control of their time, which they prefer to put elsewhere. What they are forgetting — or don’t know — is that the process has always been the goal, not the grades, and when they bypass the process — through cognitive surrender — they fail to learn.
6. If the AI does your work, then you will work for the AI — because you cannot think, reason or understand.

My one-handed conclusion is that AIs will help students learn if they are intrinsically motivated to “waste time” searching for connections and learn from their mistakes. AIs will undermine learning if students (and teachers) focus on short-term outputs (grades, graduates) — with dire consequeces for individuals and society.

AI disclosure: I wrote this post, excepting noted inserts — just as I’ve written the other 750+ posts here. Sadly, I’m not so sure about guest posts.

[3] I asked Claude Opus 4.7: I want an editorial, in 750 words, of the impacts of Trump’s tariffs on US businesses and consumers. Use data (qualitative and quantitative) from vetted sources. Explain impacts — intended and not — and outcomes that are likely to be well known before the november election.

Here’s the result in less than two minutes:

//begin

The Tariff Reckoning Comes Due

A year after “Liberation Day,” President Trump’s signature trade experiment has produced one clear winner — the U.S. Treasury, which collected roughly $187 billion more in tariff revenue in 2025 than the year before, a near-200% jump — and a lengthening list of losers. American consumers, small businesses, manufacturers, and the Republican congressional majority all have reason to worry about what the next seven months will reveal before voters deliver a verdict in November.

The premise was straightforward: punish foreign exporters, reshore factories, shrink the trade deficit, and raise revenue Trump insisted would come “from other countries.” The record tells a different story. Harvard Business School economist Alberto Cavallo, who tracks 350,000 product prices in real time, finds imported-goods prices rose about 5% since the tariffs began, while domestic goods rose 2.5% — because U.S. producers raised prices alongside their protected imports, as they always do. The Federal Reserve Bank of St. Louis concluded the duties “are already exerting measurable upward pressure on consumer prices.” Fed Chair Jerome Powell went further, attributing essentially all of 2025’s overshoot above the 2% inflation target (inflation closed the year at 2.7%) to tariffs. Contrary to the president’s claims, Americans are paying — an average of roughly $1,000 per household in 2025, per Tax Foundation estimates, with more layered on in 2026.

Businesses, for much of last year, were the shock absorbers. JPMorgan estimates firms swallowed roughly 80% of tariff costs in 2025, shielding households from the worst. That buffer is now vanishing. JPMorgan projects the business share could fall to 20% in 2026 as pre-tariff inventories deplete. The corporate pain has been considerable: General Motors booked $3.1 billion in 2025 tariff costs; Procter & Gamble absorbed roughly $1 billion and raised prices on a quarter of its SKUs; John Deere took a $300 million hit; Stellantis, McCormick, and Constellation Brands each reported meaningful drags on earnings. Economists at Yale’s Budget Lab and the Peterson Institute estimate a 12- to 18-month lag between imposition and full consumer pass-through, placing the peak squeeze between April and October 2026 — precisely the pre-election window.

The intended benefits have largely failed to materialize. Manufacturing employment, the tariffs’ totemic objective, shed 83,000 jobs during Trump’s first year, per Labor Department data. The ISM manufacturing index sank to 48.2%, signaling contraction, and its employment subindex hit 44% — factories planning further cuts. The Reshoring Institute’s own 2025 survey found one-third of U.S. equipment manufacturers still planning to move production offshore, citing cost. Headline announcements like Stellantis’s $13 billion U.S. investment commitment look impressive but, as economists including AEI’s Michael Strain and Chicago Booth’s Matt Notowidigdo argue, the policy’s chronic uncertainty has paralyzed more decisions than it has catalyzed. Meanwhile U.S. soybean exports to China collapsed 78% and corn exports 99% through August 2025 as retaliation flattened farm country.

The unintended consequences have multiplied. The Supreme Court, ruling 6-3 on February 20, 2026, held that Trump’s sweeping IEEPA tariffs were never authorized by Congress. Treasury must now refund tens of billions to importers — a logistical scramble already underway⁴ — while the administration has replaced the voided duties with a 10% Section 122 tariff that expires after 150 days and launched new Section 301 investigations against 16 trading partners. For small businesses, this whiplash has been crippling: CNBC documents owners who have cancelled expansions or closed outright. Layer on the U.S.-Israeli war with Iran, which has pushed energy and shipping costs to post-Ukraine-invasion highs, and the absorption capacity of American firms is effectively gone.

By November, three outcomes will be visible. First, voters will have experienced the sharpest tariff pass-through to shelves and showrooms precisely as ballots are cast. Second, the promised manufacturing renaissance will remain conspicuously absent; factory payrolls and the ISM index will tell their own story. Third, the political arithmetic is already materializing: 55% of registered voters tell NBC News tariffs have hurt the economy; 54% oppose them outright in New York Times/Siena polling; Democrats have pulled even with Republicans on economic competence for the first time since December 2017.

Trump’s gamble was that Americans would tolerate short-term pain for long-term revival. The data suggest they will get the first without the second — and deliver their verdict accordingly. //end

[4] I asked for more on that process: The refund process is a three-sided mess. The government loses $150–175 billion in cash it already counted, worsening a deficit it planned to close with that same tariff revenue. Businesses get a welcome but slow and uncertain liquidity injection — while facing the prospect of re-imposition under new statutory authority and a fresh class-action litigation front. Consumers, who bore much of the cost, have no direct path to recovery and are unlikely to see lower prices soon, given that the Iran conflict is simultaneously pushing energy and transport costs up. The only clear winners so far are the plaintiffs’ bar and the trade lawyers advising on CAPE filings.