Preparation changes more than presence
Split a teaching week into its tasks and the pattern is clear. Drafting lesson plans, generating practice questions, producing differentiated versions of a worksheet and summarizing resources are screen-based, structured and cheap to review, exactly the profile of work that generative AI accelerates. Managing a classroom, noticing a struggling student, de-escalating a conflict and earning trust are none of those things.
O*NET's description of secondary teaching runs far beyond content delivery: establishing rules and procedures, adapting methods to student needs, conferring with parents and guardians, and observing students for signs of difficulty. These tasks carry legal and safeguarding weight that no draft can carry. This is why task-level research treats teaching as an augmentation case: the preparation layer compresses while the human layer, if anything, grows more visible.
The new literacy: verifying and modeling AI use
Two AI skills are becoming part of the craft. The first is verification: generated lesson materials can be confidently wrong, subtly biased or misaligned with your curriculum standard, so the teacher who checks against the standard before the class sees it is doing new professional work. The second is modeling: students already use these tools, usually without guidance. A clear classroom policy on when AI use is learning and when it is avoidance teaches a judgment students will need for decades.
Neither skill requires becoming a technologist. Both extend what teachers already do: quality-control of materials, and the explicit teaching of honest work.
A realistic 90-day plan
First month: pick one recurring preparation task, such as drafting differentiated practice sets, and run it AI-assisted with a fixed verification checklist: factual accuracy, standard alignment, reading level, bias. Track the minutes saved.
Second month: spend those reclaimed minutes on the highest-value human task you never had time for, individual feedback conversations are the usual answer. Third month: write a one-page AI policy for your classroom and share it with colleagues. Teachers who can articulate a working policy become the reference point when their school inevitably needs one.
