Oregon Teachers Are Calling AI Deals a "Grift"

Not every teacher is thrilled about the industry's warm embrace of the classroom. Inside the National Education Association's Oregon affiliate, a rank-and-file caucus has been demanding an end to what it bluntly calls "AI grift" — a grassroots refusal of the enthusiasm flowing from vendors and, increasingly, from union leadership itself. It's a vivid phrase, and it's tempting to inflate it into "teachers reject AI." That would be wrong. What it actually is — one caucus, in one state affiliate, of the union that is not the one that signed the big AI deal — is more specific and, read carefully, more interesting. It names a real structural question without settling it.

Who is actually saying this

Precision first, because the framing is easy to abuse. The NEA is the country's largest teachers' union and the chief rival of the AFT, the union that brokered the marquee AI-training partnership with OpenAI, Microsoft, and Anthropic. The "grift" language is coming from a caucus — an organized faction of members — inside the NEA's Oregon affiliate. It is not NEA national policy. It is not even Oregon NEA's official position. It's the position of a rank-and-file group pushing from below, which is exactly what caucuses exist to do.

That matters for two reasons. First, it means the honest headline is "an activist faction in one state affiliate of the rival union is skeptical," not "the teaching profession has ruled against AI." Second, it tells you the pushback is internal and directional — members applying pressure to their own leadership — rather than an institutional verdict handed down from the top. Grassroots skepticism is a real and legitimate signal. It is not the same thing as a union's decision, and conflating the two would repeat precisely the kind of overreach this series has flagged elsewhere.

"AI grift" is a rhetorical frame, not a finding. It belongs in the column of attributed opinion — one caucus's view — not the column of established fact. But dismissing it on those grounds would be its own error, because the question underneath the slogan is a serious one.

The steelman of the "grift" charge

Strip away the heat in the word and there's a genuine argument, and fairness requires stating it at full strength. The concern is a conflict of interest, and it's the same one raised well beyond Oregon. When the AFT accepted funding to build a National Academy for AI Instruction — a five-year, $23 million effort with OpenAI contributing $10 million toward training some 400,000 educators — it took money from the very companies whose products it would then be training its members to adopt. Outlets including Word In Black and the Washington Informer ran pieces titled almost identically — "Teachers Union's AI Deal Raises Questions—and Concerns" — probing exactly that: is it appropriate for a union to serve as the adoption channel for the vendors funding it?

The caucus's instinct maps onto that worry. If the entity meant to represent teachers' interests is being paid by the entity selling into their classrooms, whose interest does the training actually serve? Even a scrupulously run program inherits the appearance problem, and appearance problems matter for an institution whose authority rests on member trust. "Grift" is an unfair word for a program that may be entirely above-board — but "conflict of interest worth scrutinizing" is a fair description of the structure, and the caucus is right that someone should be watching it.

The other side of the ledger

Being fair means holding the counter-argument at equal strength, and there is a real one. Unions have accepted external funding for professional development for as long as professional development has existed; that a funder benefits from a trained workforce doesn't by itself make the training corrupt. The AFT's stated design puts educators — not the vendors — in charge of the curriculum, which is the specific safeguard the conflict-of-interest critique demands. And the underlying tools carry their own protections: OpenAI's teacher offering is not used to train models by default and sits inside a FERPA-oriented data posture, which blunts the "they're just harvesting classrooms" version of the suspicion.

The broader union conversation is also more mixed than either "embrace" or "grift" captures. The NEA's own coverage of ChatGPT in schools weighs the pros and cons rather than issuing a verdict — the posture of an organization still genuinely deliberating, not one captured or in revolt. So the real landscape is a spectrum: AFT leadership leaning in via the funded academy, an NEA state caucus pushing hard the other way, and a large deliberative middle that hasn't committed. Any story that flattens that spectrum into a single mood is selling something.

Why the dissent matters even if it's a minority

Here's the thing about a caucus using a deliberately provocative word: its job isn't to be the final verdict. Its job is to force a question onto the table that leadership would rather manage quietly. On that measure, the "grift" framing is doing real work. It's a reminder that adoption is not only a product decision or a policy decision — it lands inside a contested labor politics, where members are entitled to ask who's paying for their "training" and why.

That has a concrete bearing on ChatGPT for Teachers. A free tool doesn't arrive in a neutral field; it arrives into unions that are actively arguing about how close they should stand to AI vendors, and into a workforce where a vocal minority reads the whole arrangement as capture. OpenAI can point, accurately, to teacher-designed curriculum and no-training-by-default commitments. The caucus can point, accurately, to the structural conflict of a union funded by the sellers. Both things are true at once, which is usually the sign that the honest position is "watch this closely," not "it's fine" or "it's a scam."

The fair conclusion, then, is a double one. "AI grift" overstates the case — it's one faction's slogan, not a proven charge, and the safeguards on the actual programs are more real than the slogan allows. But the question the slogan points at — can an organization take vendor money and still represent members' interests against those vendors? — is legitimate, unresolved, and worth keeping lit. The value of a minority calling something a grift isn't that they're right. It's that they keep the majority honest about a conflict the comfortable version of the story would rather not mention. That tension runs straight into the split this series turns to next: free for teachers, banned for students.

Part 70 of 100 in the ChatGPT for Teachers series. Previously: What the MIT Cognitive Engagement Study Found. Next: Free for Teachers, Banned for Students: The Split. Browse more builder insights or explore AI skills for education at aiskill.market.