The New Expectation Creep: When AI Becomes One More Thing Teachers Are Supposed to Figure Out

AI, Teachers, and the Human Work of Schools — Part 1

Artificial intelligence is arriving in public schools with the speed of a mandate and the support structure of a suggestion.

Teachers are hearing that AI may transform learning, save time, personalize instruction, streamline planning, improve feedback, and prepare students for a future already taking shape around them. Some of that promise may be real. AI may indeed become a useful tool in the daily work of schools, especially when it helps reduce clerical burdens and gives teachers more time for the relational and instructional work that matters most.

But the deeper policy question is not whether AI has potential.

The deeper question, and most importantly, is whether schools are building the professional conditions teachers need to use it wisely.

That is the focus of this three-part series: The New Expectation Creep: When AI Becomes One More Thing Teachers Are Supposed to Figure Out. Over the next three posts, I want to look beyond the excitement and anxiety surrounding AI and ask a more grounded question: What happens when a powerful new technology enters public schools faster than the systems of training, support, teacher voice, privacy protection, and professional judgment needed to guide it?

This first post begins with time.

Because before teachers can evaluate AI ethically, integrate it pedagogically, or use it in ways that protect students, they need something policy often overlooks.

They need time to learn.

The Familiar Pattern of Expectation Creep

Teachers have seen this pattern before.

A new initiative arrives with ambitious language. It is described as innovative, necessary, student-centered, future-focused, and urgent. The rollout may include a district email, a short presentation, a resource folder, a webinar link, or a list of approved tools. Sometimes there is encouragement. Sometimes there is pressure. Often, there is ambiguity.

Then the real work quietly shifts to teachers.

• Learn the platform.
• Understand the risks.
• Adapt the lesson.
• Protect student data.
• Answer parent questions.
• Monitor student use.
• Detect misuse.
• Redesign assignments.
• Stay current.
• Do it responsibly.

And somehow, do it without additional time.

This is what I have come to think of as expectation creep — the slow expansion of what teachers are expected to absorb, manage, and master without a corresponding expansion of time, staffing, compensation, or professional support.

AI is becoming the newest version of that pattern.

One teacher and education writer, Larry Ferlazzo, captured part of this tension when he wrote that AI is forcing teachers to spend “a ton of time rethinking” lessons. He was not simply rejecting AI. He was naming the hidden labor it creates. Even when a technology may eventually help teachers, there is still an implementation cost.

 Someone has to rethink the unit. Someone has to revise the assignment. Someone has to anticipate how students will use the tool. Someone has to rebuild the learning experience so it still asks students to think.

The problem is not that teachers are unwilling to learn. Teachers learn constantly. They revise, adapt, troubleshoot, research, collaborate, and reinvent parts of their practice every year. The teaching profession has always required learning on the move.

But there is a difference between professional growth and unsupported implementation.

Professional growth is structured, sustained, and connected to real practice. Unsupported implementation is what happens when a system introduces a tool, names it important, and then leaves teachers to figure it out in the margins of an already overloaded day.

The Gap Between Use and Preparation

One of the striking realities of AI in schools is that teacher use is already outpacing formal preparation.

Many teachers are experimenting with AI tools on their own. Some are using them to draft emails, create examples, generate reading passages, modify assignments, translate communications, organize lesson ideas, or reduce repetitive administrative tasks. In some cases, teachers are finding genuine relief. If a tool helps a teacher reclaim even a small portion of time in a profession shaped by constant overload, that matters.

Because teachers do not need another initiative that arrives as inspiration and becomes obligation. They do not need another set of expectations added to the edge of the day. They do not need vague encouragement to “embrace the future” without clear guidance about how to protect the present.

They need time.

They need training that is practical, sustained, and connected to the actual students in front of them. They need policies that distinguish between grade levels, subject areas, developmental readiness, academic integrity, accessibility, and data privacy. They need administrators who understand that AI is not just a tool teachers can “try out,” but a force that changes planning, assessment, feedback, student behavior, and parent communication.

But the unevenness is the issue.

Some teachers have access to training, guidance, and approved tools. Others are relying on social media tips, colleague suggestions, trial and error, or late-night experimentation. Some districts have clear policies. Others are still developing them. Some teachers know what student information can and cannot be entered into an AI system. Others are left guessing.

That is not a responsible implementation strategy.

When only a small share of teachers receive formal guidance on AI use, the result is not innovation. The result is fragmentation. Teachers are placed in the position of making ethical, instructional, and legal decisions without the shared knowledge base those decisions require.

This is especially concerning because AI is not just another digital tool. It raises questions about student privacy, bias, academic integrity, intellectual property, instructional quality, special education accommodations, grading, feedback, and the future role of professional judgment in schools.

Those are not small questions.

They should not be answered by individual teachers in isolation at 9:30 at night after the rest of the day’s work is done.

Time Is a Policy Issue

In education policy, time is often treated as a scheduling matter.

But time is also a justice issue. It is a workload issue. It is a professional respect issue.

When policymakers or district leaders say that teachers need to become AI literate, they are naming a legitimate need. But if they do not also provide protected time for that learning, then the policy message becomes something else: this matters enough for teachers to be responsible for it, but not enough for the system to make room for it.

That is where expectation creep becomes harmful.

Teachers are already managing curriculum demands, assessment requirements, student needs, family communication, behavioral concerns, documentation, professional learning communities, special education paperwork, intervention plans, data meetings, and the emotional labor of caring for children in a complicated time.

AI does not enter an empty room.

It enters a crowded one.

So when AI is added without time, it does not feel like support. It feels like one more thing teachers are expected to carry.

This is where public policy and district policy need to be clearer. If AI literacy is now part of the professional landscape of teaching, then AI learning must be built into the professional structure of teaching. It cannot depend primarily on personal curiosity, unpaid labor, or the energy teachers have left after everything else has been done.

The Risk of “Incidental Learning”

There is value in informal exploration. Teachers have always learned from one another, shared shortcuts, passed along resources, and discovered practical strategies through experience. Some of the best professional knowledge in schools comes from hallway conversations, department meetings, and one teacher saying to another, “Here’s what worked for me.”

But incidental learning cannot be the foundation of AI policy.

AI tools are changing too quickly, and the stakes are too high. Without formal guidance, teachers may use tools that have not been vetted for privacy. They may unknowingly enter protected student information into systems that store or reuse data. They may rely on AI-generated materials that contain errors, bias, weak pedagogy, or fabricated information. They may receive conflicting messages about whether AI use is encouraged, discouraged, required, or prohibited.

Even more importantly, incidental learning tends to deepen inequity.

Teachers with more time, stronger technology backgrounds, better district support, or access to active professional networks may become confident AI users. Teachers in under-resourced schools, or those already carrying heavier workloads, may be left further behind. The result is a new kind of digital divide — not only among students, but among educators.

If AI is going to be part of public education, access to AI professional learning cannot depend on a teacher’s personal bandwidth.

It has to be part of the system.

What Better Policy Would Require

A humane AI policy begins by acknowledging that teachers need more than access to tools. They need structured time to understand those tools, evaluate them, question them, practice with them, and decide where they do and do not belong.

That means AI professional learning should be built into contract time, not added as optional evening webinars or self-paced modules teachers are expected to complete on their own. It should be ongoing rather than one-and-done. It should be differentiated by grade level, subject area, student population, and instructional purpose. A kindergarten teacher, a high school English teacher, a special education teacher, and a school counselor are not all asking the same AI questions.

They should not receive the same generic training.

Good policy would also distinguish between technical training and professional learning. Teachers do not only need to know how to write a prompt or navigate a platform. They need to understand when AI improves learning, when it weakens learning, when it saves time, when it creates risk, and when it interferes with the human relationship at the center of teaching.

That requires more than a tutorial.

It requires professional conversation.

Districts should create space for teachers to bring real problems of practice: How might AI help me modify a reading passage without lowering expectations? How do I teach students to use AI without outsourcing their thinking? What should I never enter into a tool? How do I respond when students submit AI-generated work? How do I use AI to reduce clerical load without allowing it to shape my professional judgment?

These are not questions that can be solved by a vendor demonstration, A webinar, a PowerPoint presentation, or classes viewed online.

They are questions that require teachers in the room.

Teacher Voice Belongs at the Center

One of the mistakes schools often make with technology initiatives is confusing rollout with implementation.

A rollout can happen quickly. Implementation takes trust.

If teachers are not part of shaping AI policy, the result will likely be either overuse, underuse, confusion, or resistance. Not because teachers oppose innovation, but because they know what happens when tools are introduced without attention to classroom reality.

Teachers know which tasks are genuinely time-consuming. They know where students struggle. They know which forms of feedback matter. They know when a tool is helping and when it is creating more work than it saves. They know that what looks efficient from a district office may feel very different in a classroom full of actual students.

For that reason, every district AI policy should include teacher voice from the beginning. Not after the purchase. Not after the platform is selected. Not after the expectations are written.

Before.

Teacher advisory groups, curriculum committees, union-management conversations, pilot teams, and building-based feedback loops should all be part of AI decision-making. The teachers who will be expected to carry the work need a meaningful role in defining the work.

This is not simply a matter of morale.

It is a matter of policy quality.

Policies designed without teacher voice often underestimate complexity. Policies designed with teacher voice are more likely to be realistic, ethical, and usable.

The Question Behind the Tool

AI may eventually help teachers save time. That possibility should not be dismissed. In a profession where time is one of the scarcest resources, any tool that reduces unnecessary workload deserves careful attention.

But there is an important difference between a tool that saves time and a policy that shifts responsibility.

If teachers are expected to learn AI on their own, vet tools on their own, manage risks on their own, and redesign instruction on their own, then AI has not reduced workload. It has simply relocated it.

The question is not only, “Can AI help teachers?”

The question is, “Will schools create the conditions that allow AI to help without adding another layer of invisible labor?”

That question should sit at the center of every AI policy conversation.

Because the future of AI in schools will not be determined by the tools alone. It will be determined by the human systems built around them.

A More Humane Starting Point

A more humane AI policy would begin with a simple acknowledgment:

Teachers should not have to self-teach their way into the future of education.

If AI literacy matters, fund it.

If responsible use matters, teach it.

If privacy matters, clarify it.

If innovation matters, give teachers time to practice.

If student learning matters, keep professional judgment at the center.

Public schools do not need another initiative that depends on the quiet exhaustion of teachers. They need policies that recognize the human cost of implementation and make room for the learning that responsible change requires.

AI may become part of the future of teaching.

But if that future is going to be wise, ethical, and humane, it cannot be built on expectation creep.

It has to be built on time, trust, training, and teacher voice.