AI needs clearer rules
Schools are seeing a gap between student access to generative AI and formal guidance on how to use it responsibly. A practical guide finds many students use AI but only 42% have received ethical-use guidance, UNESCO urges transparency and ethical framing, and a Scientific Reports paper flags both benefits for tutoring and the need for transparency in AI systems (nerdbot.com; unesco.org; nature.com).
Students are using generative artificial intelligence in school faster than schools are writing rules for it. Stanford’s 2026 AI Index says half of middle and high schools have artificial intelligence policies, but just 6% of teachers call those policies clear. (hai.stanford.edu) Generative artificial intelligence is software that predicts the next word, image, or answer from patterns in large datasets. In classrooms, that now means students use it for research, essay editing, and brainstorming, according to the Stanford Human-Centered Artificial Intelligence 2026 AI Index. (hai.stanford.edu) The policy gap shows up in guidance as much as access. A practical guide published April 14 by NerdBot said many students already use artificial intelligence tools, but only 42% reported receiving ethical-use guidance from their schools. (nerdbot.com) The United Nations Educational, Scientific and Cultural Organization has been warning that education policy is lagging behind the technology. On its artificial intelligence and education page, updated April 8, UNESCO said rapid advances have outpaced policy debates and regulatory frameworks. (unesco.org) UNESCO has spent the past few years building the policy scaffolding schools are now being pushed to use. It points governments to its 2019 Beijing Consensus, its 2021 policy-maker guidance, and its competency frameworks for students and teachers as the baseline for human-centered, equitable use. (unesco.org) Researchers are also trying to define what “responsible use” looks like in practice. A Scientific Reports paper published April 13 described a “transparency score” on a 0 to 100 scale for whether an artificial intelligence system’s decisions are explainable, traceable, and accountable to students, teachers, and administrators. (nature.com) That paper, based on data from 2,847 students across three Pakistani universities, argued that personalized learning gains come with risks around privacy, bias, and opaque systems. The authors said schools need clearer disclosure about how tools work and how decisions are made. (nature.com) The case for using artificial intelligence in teaching is not hypothetical. A separate Scientific Reports randomized controlled trial published June 3, 2025 found college students learned significantly more in less time with an artificial intelligence tutor than with an in-class active-learning lesson, while also reporting higher engagement and motivation. (nature.com) But that same 2025 paper also drew a line between guided tutors and unguided chatbots. The authors said general-purpose systems can let students finish assignments without doing the thinking, and can answer confidently when they are wrong. (nature.com) Another April 2026 study in AI and Society found the same transparency problem on the product side. After reviewing 20 generative artificial intelligence tools marketed to teachers, the researcher reported that most did not clearly explain how they worked, and none disclosed training procedures or data sources. (link.springer.com) Schools are now in the middle of those two facts: students already have the tools, and the rules are still being drafted. The immediate fight is less about whether artificial intelligence enters the classroom than about who explains it, when, and under what terms. (hai.stanford.edu)
