Why a Union, Not a Vendor, Leads Teacher AI Training

Here is a genuinely odd sentence: OpenAI, Microsoft, and Anthropic gave $23 million to a teachers' union so the union could train teachers on AI. Not so the companies could. The union. If you find that a little suspicious, you are not alone — and you are asking exactly the right question. Why would companies pay someone else to do the very thing they are perfectly capable of doing themselves, and would presumably prefer to control?

The answer is where this whole cluster of articles has been heading. It turns on a single scarce resource that no amount of money can manufacture from the inside: trust. But the skeptical reading — that this is a sophisticated influence play dressed up as public service — deserves a real hearing, not a wave of the hand. This article gives the skeptics their due, names their strongest case, and then makes the argument for why a union in the lead still matters.

Key Takeaways

• Vendors cannot self-issue credibility. When a company teaches you to use its own product, part of you files it as sales. A union teaching the same material carries trust a vendor cannot buy.
• The skeptics have a point worth naming. Futurism argued the funding looks like an influence play — a way for AI firms to shape how teachers think about AI. That tension is real and should not be dismissed.
• Neutrality is the entire value. The AFT can vet, refuse, and endorse in a way no funder can. Multi-funder backing (not one company) is what makes that independence structurally possible.
• Follow-the-money and public-good can both be true. The companies benefit from a confident market and teachers get union-shaped training. Motive and value are not mutually exclusive.
• The test is governance, not intentions. Whether the union stays in the driver's seat — controlling curriculum and keeping it tool-agnostic — is what separates real independence from a branded funnel.

Why a Vendor Simply Can't Lead This

Start with the structural problem. Imagine OpenAI running the national teacher-AI curriculum directly. Every lesson, no matter how honest, arrives with an asterisk: this is the company whose product you might buy, teaching you why AI is great. That is not a knock on OpenAI specifically — it is true of any vendor teaching about its own category. The credibility tax is unavoidable, and in education, where skepticism about corporate influence runs deep and appropriately so, that tax is steep.

A union does not pay it. When the AFT tells its members "we vetted this training and it serves your interests," teachers extend a trust that is built over decades of the union representing them against, among others, exactly these kinds of outside interests. The AFT's own framing positions the effort as educator-led. That leadership is the product. It is why the multi-funder structure matters so much: money from three competitors, routed through a neutral body, is far harder to read as any single company's marketing than money from one. The coopetition logic behind that structure is the subject of Why OpenAI, Microsoft, and Anthropic Fund One Union.

The Skeptic's Case, Given a Fair Hearing

Now the counter-argument, and it is a serious one. Futurism published a pointed piece — headlined, roughly, that you'll never guess why OpenAI is pouring money into the second-largest teachers' union — questioning the motives behind the funding. You can read it directly at Futurism. The core of the skeptical view goes like this: companies do not spend $23 million out of altruism. If you fund the institution that shapes how 400,000 teachers understand AI, you are buying influence over the frame — what gets taught as normal, what gets treated as inevitable, whose tools become the default reference. A neutral-looking union is, on this reading, the most effective possible vehicle for that influence because it doesn't look like influence.

This is not a paranoid take, and it should not be dismissed as one. It rhymes with a long history of industry funding shaping the terms of debate in fields from nutrition to energy. A teacher would be right to ask: who wrote the curriculum, who can veto it, and what happens if the union's independent judgment collides with a funder's interest? Those are the questions that separate genuine independence from a well-disguised sales funnel, and they connect to the broader "why give it away" scrutiny this series takes up in Follow the Money: Why OpenAI Gives ChatGPT Away and The Skeptic's Case Against Free AI From OpenAI.

Can both things be true at once?

Yes — and this is the part the loudest versions of both sides miss. It can be simultaneously true that the companies fund this because a confident, AI-fluent teacher market serves their business, and that teachers get better, more independent training than any vendor would provide alone. Motive and value are not opposites. As TIME and EdWeek both noted, the arrangement is unusual precisely because it puts a self-interested motive and a public-good outcome in the same structure. The interesting question is not "which one is real" — both are — but "which one wins when they conflict."

What Actually Determines the Outcome

The honest conclusion is that intentions are unmeasurable and governance is not. The safeguard against the skeptic's fear is not the companies' good faith; it is the union's control. If the AFT owns the curriculum, keeps it tool-agnostic, and is willing to tell a funder no, then union leadership means what it claims to mean. If the training quietly becomes a tour of one company's features, the skeptics were right. That is an observable outcome, and it is worth watching for.

For a teacher, the practical defense is fluency itself. A teacher who can independently evaluate AI output — the core skill defined in What "AI Fluency" Actually Means for a Teacher — is exactly the person least susceptible to any vendor's framing, whoever funds the training. The best insurance against influence is a workforce that thinks critically about the tools, which is precisely what a genuinely independent, union-led program should produce. Building that instinct on real work — with tool-agnostic education agent skills or a study habit coach you can test against your own judgment — is the same muscle, whether the tool is ChatGPT or the parallel effort Anthropic runs in the Claude for Teachers series. Keep the union in the driver's seat, keep teachers thinking critically, and the arrangement earns its trust. Let either slip, and it doesn't.

Frequently Asked Questions

Why doesn't OpenAI just train teachers directly?

Because a vendor teaching about its own product carries an unavoidable credibility tax — teachers file it as sales. A union can vet and endorse training with a trust no company can self-issue, which is the entire value of the arrangement.

What is the skeptical view of the funding?

Futurism argued the money looks like an influence play — a way for AI companies to shape how teachers understand AI by funding the institution that frames it. It is a serious argument rooted in the history of industry-funded education, not a fringe one.

Is the skeptical view correct?

It captures a real risk. But motive and value can coexist: the companies can benefit and teachers can get better training than any vendor would give alone. The outcome depends on governance — whether the union genuinely controls the curriculum.

How would you tell if the union is really independent?

Watch whether the curriculum stays tool-agnostic and whether the AFT can say no to a funder. Independence is proven by governance and outputs, not by anyone's stated intentions.

What protects teachers regardless of who's right?

Fluency. A teacher who can independently evaluate AI output is the least susceptible to any vendor's framing. A genuinely independent, union-led program should produce exactly that kind of critical thinker.

Part 20 of 100 in the ChatGPT for Teachers series. Previously: What "AI Fluency" Actually Means for a Teacher. Next: ChatGPT for Teachers vs. Gemini for Education. Browse more builder insights or explore AI skills for education at aiskill.market.