I taught English in an asynchronous online setting when AI submissions started flooding in. At first it was a trickle. A few papers that felt off, word choices that did not match a student’s previous work, arguments that were strangely polished and utterly hollow. Then it became a wave. Then a tidal wave.
Suddenly my job had changed. I was no longer just teaching writing. I was documenting suspicions. Adjusting feedback to probe for understanding. Finding ways to approach students about potential AI use without being accusatory. The moment students felt called out, I was met with resistance, denial, and very often, angry emails from parents demanding to know why I was accusing their child of cheating.
I had to figure all of this out on my own. And so does every other teacher.
The Work That No One Talks About
Administrators and policymakers do not seem to understand that AI has changed the nature of teaching, not only what students submit. And that change has come with an enormous amount of invisible labor that teachers are expected to absorb without acknowledgment, without support, and without compensation.
AI capabilities change constantly. What Claude, and Gemini could do eighteen months ago is nothing compared to what they do now. Teachers don’t have time to stay current on best practices and research. Most of us are doing this on our own time, reading articles and experimenting after hours, because there is no time built into our contracts for this kind of ongoing education. Often, a technology integrationist is sent to a conference or a workshop and then tasked with coming back to train the entire staff as if they are now the expert. This is not an effective method of professional development that supports educators. It’s a high stakes game of telephone (remember those?!)
Every assignment I had refined over decades of teaching was suddenly suspect. Could students complete it with AI? (Probably) Would AI-generated work be distinguishable from student work? (Not for long) What would make an assignment “AI-resistant” and is that even possible? (It’s getting harder) That meant rebuilding curriculum from the ground up while still teaching with the broken version.
Evaluating whether a submission is AI-generated takes time. Far more time than grading authentic student work! And when I suspect AI use, I face an impossible choice: say nothing and let it go, or initiate a conversation that may result in conflict, parental complaints, and hours of additional documentation and meetings. Either way costs something. And I never let anything go 😉
One Person Cannot Hold Back the Tidal Wave
I tried. I documented. I adjusted. I had careful conversations. I revised my assignments. I read everything I could find about AI detection and its limitations. I developed approaches that felt fair and pedagogically sound.
And I watched colleagues give up.
I do not blame them. I considered it myself. Without proper support and guidance, some teachers simply are not trying to fight it anymore. They are grading AI submissions and giving credit because the alternative (the confrontation, the documentation, the parent emails, the administrative meetings) is not worth the cost. They are exhausted. They were exhausted before AI arrived. Now feeling like human plagiarism detectors, we fail to find the same meaning we once did in asynchronous online education.
This creates a deeply inconsistent experience for students. In one class, AI use might result in a zero and an academic integrity investigation. In another class, it might result in full credit and no questions asked. Students learn quickly which teachers check and which teachers do not. The message we send is that academic integrity depends on who catches you, not on any consistent standard.
This has always been somewhat true. Teachers have always varied in their grading and their vigilance. But the stakes are different now. We are talking about students potentially earning credits for work they did not do, developing none of the skills the coursework was designed to build. Now the variation between teachers decides whether learning happens at all.
The Inequity No One Mentions
This burden does not fall equally on all teachers.
Teachers in under-resourced schools, without professional development budgets, without adequate planning time, are expected to figure this out with fewer resources than their counterparts in wealthy districts. They have the same challenges and fewer tools to address them.
Not everyone has the time or interest to dive deep into AI topics. I do not think that is a moral failing. Teachers are already working unsustainable hours. Asking them to become AI experts on top of everything else they do is not reasonable. But the result is that some teachers are adapting and some are not, and students experience wildly different educational environments depending on who happens to be at the front of their classroom.
What Needs to Change
This is not sustainable. Something has to give, and right now what is giving is teacher wellbeing and educational quality. Here is what I think needs to happen:
For districts and policymakers:
Stop pretending this is a minor adjustment. AI has fundamentally changed the teaching profession, and responding to that change requires resources: time, training, and support structures that most districts have not provided. Teachers need dedicated professional development that goes beyond a single workshop. They need ongoing support from experts who understand both pedagogy and AI. They need policies that are clear, consistent, and actually enforceable.
Recognize the inequity. Schools with fewer resources need more support, not less. If we do not address this, we will create a two-tiered system where students in wealthy districts get teachers who are equipped to handle AI and students in under-resourced districts get teachers who are drowning.
Build time into contracts. Curriculum redesign is real work. Learning new technology is real work. Developing AI-aware assessment strategies is real work. If you expect teachers to do this work, you need to pay them for it and give them time to do it. Expecting it to happen on evenings and weekends is exploitation.
For teachers navigating this now:
If this feels impossible, that’s because it is, at least in the way we are currently being asked to do it. The expectation that individual teachers can solve a systemic problem through individual effort is unreasonable.
Find your people. Connect with other teachers who are grappling with the same challenges. Share what works. Admit what does not. The isolation makes this harder than it needs to be.
Protect yourself. Document your efforts. Keep records of the time you spend on AI-related work. When you advocate for more support, having data helps. And when you are exhausted and need to let something go, let it go. Your health matters more than catching every AI submission.
Be honest about your limits. You cannot do everything. Deciding where to focus your limited energy is how you survive this. Only you can decide which battles are worth fighting.
The Tidal Wave Is Still Coming
AI is not going away. It is going to get better, more accessible, and more integrated into students’ lives. The challenges I faced three years ago are nothing compared to what teachers face now, and what teachers face now will seem quaint in another two years.
We can keep asking individual teachers to hold back the tidal wave with their bare hands. We can keep pretending that willpower and dedication are sufficient responses to technological disruption. We can keep burning through educators until there are none left willing to do this work.
Or we can acknowledge that the system is broken and start building something that actually supports the people we expect to educate the next generation.
I know which option I would choose. I am still waiting to see which option our institutions will choose.