OpenAI announced Education Plugins today, built separately for K-12 educators, college educators, and college students. The plugins are bundles of apps, instructions, role-specific skills, and recurring workflows, moving AI from producing answers on request to carrying out multi-step work that a student or teacher must supervise and judge. The design follows OpenAI's distinction between ChatGPT Chat, which handles questions and short tasks, and ChatGPT Work, which is built for projects involving multiple sources, tools, and deliverables. Work proposes a plan, names missing context, completes approved steps, and stops when it needs guidance. Freeing users from composing elaborate prompts increases the demand on human ability, as users must define the objective, choose sources, set constraints, decide what an acceptable result looks like, catch errors, and know which decisions should never be handed off. The distinction matters for students because school and work treat finished products differently; a student can produce a competent report without understanding it well enough to defend it. Banning the tools would leave students unready for workplaces where these systems are ordinary equipment, while accepting AI-produced work without changing assignments would make it harder to know what anyone has learned.