*October 2nd 2026* This is a shorter companion to my longer essay [[Avoiding Death by Meat-Proxy|Avoiding Death by Meat-Proxy: Moral Ownership and AI in Schools]]. A problem in schools at the moment is the abundance of AI-slop: presentations, posters, and policies. Students notice, and they hate it. That said, there are legitimate uses for AI in schools. In fact, I use it almost daily. The question is not a binary *to AI or not to AI*. Instead, the question is **when does using AI become a problem?** For much of our work as teachers, **moral ownership** is vital. Moral ownership comes when we do the work of thinking, and our creations become part of us. It comes when we inhabit them and they inhabit us. When we communicate—with students, each other, and parents—it matters that we have moral ownership of what we're saying. Moral ownership matters because it gives legitimacy to our message and integrity to our voice. Often, our voice is fundamental to our message as it is heard in the context of our relationships. As teachers, our relationships are underpinned by trust and anything that undermines this trust should be anathema. It's not important that we have moral ownership of everything we do—we don't need to hand-craft every worksheet ourselves—but it's critical that we take seriously the responsibility we have to do right by our students, and that we are able to account for the choices we make in our teaching. We need to understand (and be able to defend) our pedagogical choices because of the significance of their impact. For much of what we do, we must do the **work of thinking** ourselves rather than outsource it. AI becomes a problem when it starts to 'think' for us about things we should have moral ownership over. A practical, though imperfect, approach is to consider the primary purpose of whatever we are working on. **Functional** work's value comes from what it does. For example, worksheets, spreadsheets, and practice questions succeed on their quality rather than their provenance. AI can help produce functional things like this—in fact, it can do it faster than we can, and free us up to think about the big picture of our teaching rather than the minutiae of a worksheet's wording. We are still, however, responsible to our students so must ensure quality, as we should with any resource. **Communicative** work is different. When we communicate, human connection matters. The fact that I am speaking to you is meaningful. Our voices matter because they are part of our message. Emailing colleagues and parents, delivering assemblies, and presentations to staff gain value (hopefully) from the voice, credibility, and integrity of the person speaking. Letting AI speak for you robs you of your voice and diminishes your message, turning you into a meat-proxy. The distinction between functional and communicative work breaks down fairly quickly when we consider the signals students read into things we think are purely functional. If students think we have shirked our duty and intentionally given them slop, the trust that's fundamental to our relationship is shaken. So we need to make sure we aren't giving them slop. We also must be honest about when we have used AI. As teachers, we have a moral duty to model ethical practice. Ethical practice includes not claiming ownership of things that aren't ours either explicitly or by omission. This means we must disclose when resources we're using are AI-generated. More broadly, we should disclose when we have used AI because silence implies human ownership and this is dishonest. Determining moral ownership is messy. When I use AI to make software for me, I feel like I own a lot of it because I do lots of the work of thinking. However, I don't do all the thinking as AI handles the implementation for me. I don't need to understand every line of code to have a meaningful stake in the product, but I must accept that my stake is partial—and this is okay. I must be transparent about the fact the work is not entirely mine because doing otherwise undermines the trust fundamental to my relationships. AI should not displace the important human elements of our work. As teachers, much of what we do succeeds because of its human elements. AI can help us make things and reduce tedious tasks, but should not become a substitute for our professional judgement. Before using AI, you should ask two questions to determine whether moral ownership matters: - Does it matter who is speaking? - Does it matter that *I* make this? If 'yes', you should do the thinking yourself. If 'no', go ham. And disclose it. Be aware that whatever AI makes is your responsibility. You may one day have to account for it, and you can't if you were absent from the work of thinking. ## AI disclosure I drafted and edited this with the help of ChatGPT and Claude.