What Eighty-Two Percent Are Doing Alone

On May 27, Gallup and the Walton Family Foundation released the latest segment of their Teaching for Tomorrow study, a survey of 2,069 American public school teachers conducted in the early weeks of this year.¹ The number that traveled was eighty-two. Eighty-two percent of teachers said they have received no formal guidance from their school or district on how to apply artificial intelligence to their work, across the ten different tasks the study asked about.¹

The numbers underneath the headline are more specific. Sixty-nine percent of teachers reported receiving no guidance, formal or informal, on using AI for one-on-one instruction or tutoring.² Fifty-eight percent had no guidance on AI for grading or student feedback.² Forty-seven percent had no guidance on creating assignments and class materials.² About a third, thirty-four percent, had received no guidance whatsoever about AI in their classroom.¹

At the same time, three in ten teachers in the study say they are now using AI at least weekly, and that those teachers report saving an average of 5.9 hours a week.³ Six weeks of recovered time across a school year. Roughly the length of a summer session. A meaningful gift, returned to people whose hours are not easily replaced.

Read alongside each other, the two findings describe something the press release does not name out loud. The tool arrived before the conversation about what the tool is for.

What the Quiet Was Filled With

Most teachers who started using AI did not wait for their school to tell them how. They sat down at the end of a long day, opened a tab, and discovered, gratefully, that a draft rubric or a parent letter or a science worksheet could come back in two minutes instead of forty. The relief was real. The savings were measurable. The decisions about what crossed the line, and what should never be touched, were made by each teacher alone, in the dark, with no one to ask.

That is not a critique of the teachers. It is a description of what happens when a profession is handed a transformative tool and not handed a doctrine alongside it. The Gallup report shows that informal guidance, a chat in the staff room, a Slack channel, a curious principal, was offered to about half of teachers.¹ Formal guidance, the kind written down and shared as policy, reached fewer than one in five.¹ And it reached unevenly. Teachers in higher-income schools were noticeably more likely to receive any guidance at all.²

The shape of that disparity matters. In the schools with the most resources, the conversation about AI has begun, halting and uneven though it is. In the schools with the least, teachers are deciding on their own what counts as a useful shortcut, what counts as outsourcing the relationship, and what their students should be allowed to do. Whatever beliefs that conversation eventually settles into, they are being formed now, alone, at kitchen tables after the dishes are cleared.

The Conversation the Tool Was Waiting For

What the future of education needs in the next year is not, mainly, more AI tools. It is the conversation the tools should have arrived with. A teacher does not need a policy document to know that using AI to draft a parent email is different from using AI to write a student's individualized plan. She needs a school that has thought, together, about which side of the line each of those falls on. She needs to know that a colleague three classrooms down has thought about the same question. She needs to know that the principal will back her when she explains it to a parent.

That conversation is local. It will not be solved by a federal framework or a vendor's compliance page. It will be solved by school leaders sitting down with their teachers and looking carefully at what is actually happening in the work. Which is where the harder problem starts. Most of the work that AI has changed is invisible. The lesson got written. The parent got the email. The draft came back marked. The teacher saved time. The school saw none of it.

This is one of the reasons Koan has been built the way it has. Not as a tutor that replaces the conversation. As a way of preserving the trail of practice, so the conversation can finally happen. The drafts a student kept and the revisions she made. The prompts a teacher chose and the changes she made to the AI's output before handing it back. The pauses, the second tries, the small judgments that today vanish behind the polished result. Without that trail, school leaders trying to write a thoughtful AI policy are working from anecdote. With it, they can write from evidence.

Eighty-two percent of American teachers are using a new technology without a map.⁴ The map is not, fundamentally, a document. It is a habit of looking at the work, together, with care. Until that habit returns, the hours AI gives back will not become the time that schools needed most.

If your teachers got six weeks of their year back this spring, what did your school choose to do with that time?