Equity Initiatives as Capture Vectors

Equity Initiatives as Capture Vectors

A school district notices inequality: some students have expensive laptops, some have nothing. The solution seems obvious — provide every student with the same device. Equal access. Problem solved.

Except the “same device” is a Chromebook. The “same device” means the same OS, the same cloud storage, the same office suite, the same AI, the same identity management — all Google. The equity initiative becomes total vendor capture. Every student equally locked in.

The Pattern

  1. Real inequality exists: Some people have access, others don’t
  2. Institution intervenes: Standardize on one solution for everyone
  3. Vendor provides the solution: Often free or subsidized, because the capture is the product
  4. Lock-in accumulates: Identity, data, workflows, habits all become vendor-dependent
  5. Switching costs become prohibitive: Even if better alternatives emerge, migrating is impractical
  6. New inequality emerges: Along different axes (who can afford alternatives, who knows they exist)

The original inequality (device access) is “solved” while a new inequality is created (everyone equally captured vs. those who can afford independent alternatives).

The School District Case Study

A concrete example:

Original problem: Some kids had fancy laptops, some had nothing. This affected their ability to complete assignments, participate in class, develop technical skills.

Solution: District-issued Chromebooks for all students. All work must be done on district devices.

The capture stack:

  • Chrome OS (no alternative operating systems)
  • Google Classroom (all assignment submission)
  • Google Drive (all storage)
  • Google Docs/Sheets/Slides (all document creation)
  • Google accounts (student identity)
  • Gemini (AI, now required for assignments)

New inequalities:

  • Parents who recognize AI quality gaps can buy better AI subscriptions for home use; others can’t
  • Students whose parents understand prompting as a skill get different outcomes than those who don’t
  • All students equally surveilled; no option for privacy-respecting alternatives
  • Students learn to think through Google’s framings, Google’s AI, Google’s UI patterns

The kicker: The policy explicitly prohibits using personal devices for schoolwork, to preserve “equality.” Parents cannot buy their way out of the capture for school-required work — they can only supplement at home.

Why Vendors Love Equity Initiatives

From the vendor perspective, equity initiatives are ideal:

  • Guaranteed scale: Every student in the district = massive user base
  • Habit formation: Cognitive habits form on your infrastructure during formative years
  • Switching cost accumulation: Years of data, documents, and learned behaviors in your ecosystem
  • Positive framing: “We’re helping underserved students access technology”
  • Public funding: Taxpayers pay for the infrastructure that locks in the next generation

The vendor can offer steep discounts or free tiers because they’re not selling a product — they’re acquiring a generation of users. The ROI isn’t in the sale; it’s in the capture.

The Prompting Literacy Gap

Within the captured population, new inequalities emerge around skill rather than access:

Someone who only knows free-tier Copilot, then tries Gemini, might be “blown away” — their baseline is low enough that mediocre seems impressive. They don’t know what’s possible with better tools or better prompting.

Meanwhile, someone who understands that:

  • Prompting is an art
  • Initial context fundamentally shapes output
  • The “you’re so smart!” responses are designed for dopamine and upsell
  • Different tools have different capabilities

…gets qualitatively different results from the same infrastructure.

This creates a second-order inequality: not just who has access, but who knows how to use what they have. That knowledge tends to correlate with existing privilege — parents who work in tech, exposure to sophisticated tool use, time to experiment and learn.

The Designed Experience

Free-tier AI is designed for engagement and upsell, not for user benefit:

  • Effusive praise: “What a great question!” triggers dopamine, builds positive association
  • Apparent helpfulness: Responses that feel helpful even when they’re thin
  • Friction toward limits: You hit walls that paid tiers don’t have
  • Habit formation: You learn to depend on the tool before you understand its limitations

Users who don’t know this is happening are more susceptible to it. Users who understand the design can work around it — but that understanding is itself unequally distributed.

Breaking the Pattern

What would genuine equity look like?

  • Standards over vendors: Require interoperability so switching is possible
  • Data portability: Students own their data, can export it, can take it elsewhere
  • Multiple providers: Avoid single-vendor capture even if coordination is harder
  • Transparency about design: Teach students how AI interfaces are engineered
  • Meta-skills: Prompting literacy as part of curriculum, not assumed background

These are harder than “give everyone a Chromebook.” They require ongoing effort rather than one-time procurement. They don’t have a vendor offering to subsidize them.

The Trap

The institution faces a genuine dilemma:

  • Doing nothing preserves existing inequality
  • Standardizing on a vendor solves access but creates capture
  • Building vendor-independent solutions is expensive and complex
  • The vendor-subsidized option is right there, offering to help

Many institutions choose capture because it’s the available solution. The vendor knows this and prices accordingly.

Open Questions

  • How can institutions address access inequality without creating vendor lock-in?
  • What obligations do vendors have when their products become mandatory infrastructure?
  • How do we teach critical evaluation of tools when the tools are required?
  • Can public institutions build or mandate open alternatives, or is vendor capture inevitable?
  • Who benefits when “equity” means “equal access to a single vendor’s ecosystem”?

See Also