AI in Schools and the Economy: Why Adoption, Forecasts, and Spending Need Scrutiny
Artificial intelligence is often presented as an economic and educational inevitability. But recent reports point to a more complicated reality: decisions about AI are shaped not only by what the technology can do, but also by who controls access, who pays for adoption and how its benefits are measured.
A report by *The New York Times* says Google plans to turn off Search for schools that reject AI. If implemented as described, the move would raise a central question for public institutions: can schools choose whether to adopt new AI tools without risking access to other digital services? Search and generative AI are not the same thing, and treating them as a bundled choice could make it harder for educators to set policies that reflect their students’ needs.
Schools need room to make informed choices
AI can support learning through tutoring, translation and help with routine tasks. It can also produce inaccurate answers, expose sensitive information or make it harder to tell whether students understand a subject. Those risks do not mean schools should reject AI outright. They do mean adoption should involve clear rules, staff training and age-appropriate safeguards.
Schools also need meaningful alternatives. When a technology provider has a dominant role in classroom infrastructure, a decision about one product can affect access to others. Educators and families should be able to understand what data a tool collects, how its outputs are used and what happens if a school decides not to use it. A genuine choice is especially important when students cannot independently opt out of the systems their school provides.
Economic forecasts deserve scrutiny
The financial case for AI can sound just as sweeping. *WIRED* recently reported that a widely cited projection of a £400 billion boost to the UK economy was based mainly on estimates from Google. That does not automatically invalidate the forecast: companies developing AI may have useful data and expertise. But a projection is not a guarantee, and its source matters.
Economic estimates depend on assumptions about how quickly businesses adopt AI, which jobs and tasks change, and whether productivity gains translate into higher wages, lower prices or increased profits. A large headline figure can obscure those uncertainties. Policymakers evaluating AI should compare independent research, publish the assumptions behind their forecasts and track results over time.
The practical question is not simply how much value AI might create. It is who receives that value, who bears the costs and whether workers and communities have a say in the transition. Productivity gains may be real, but they are not automatically shared.
Spending discipline is part of AI strategy
A Reuters report, citing Bloomberg News, says Apple is cutting employee AI budgets as part of a cost-control effort. The reported move offers a reminder that even large technology companies face limits on how quickly they can expand AI spending. Investment is not proof of success, and retrenchment does not necessarily signal that a technology has failed. Both can reflect a search for uses that deliver measurable results.
For organizations adopting AI, the lesson is to start with specific problems rather than broad promises. A pilot should have a clear purpose, defined costs and a way to assess accuracy, safety and time saved. If a tool does not improve work enough to justify its expense—or creates new risks—teams should be able to revise or stop the project.
AI’s future will be shaped as much by procurement, policy and accountability as by technical breakthroughs. The most durable approach is neither automatic adoption nor blanket rejection. It is careful evaluation, transparent evidence and the freedom to choose tools that serve people’s needs.