Here is a hypothetical. A high school English teacher is a few weeks into a new year. Her district adopted AI guidance over the summer, and she has read it carefully. The guidance names two approved tools, asks teachers to disclose when AI has shaped feedback or materials, and prohibits entering student names or identifying details. She has a stack of essays and three days to return them, so she removes the names, runs the essays through one of the approved tools, reads over what comes back, and adds a disclosure line to each comment. Every step is within the policy. One of the essays is about a student’s grandfather, who died over the summer. The feedback on it is polite and accurate, with notes on the thesis, the transitions, and the conclusion. She returns it with the others.
Nothing in that hypothetical breaks a rule. I am not sure that settles whether she did right by that student, and I suspect she would not be sure either. What interests me is that the policy has no way to answer the question, and was never meant to.
I should say up front that I help write documents like the one she followed. Part of my work involves helping districts think through AI guidance, and I believe that work matters. A district without guidance leaves teachers guessing, and the guessing tends to fall hardest on the teachers with the least support. Still, I keep running into the same worry. The better the guidance gets, the easier it becomes to treat following it as the whole of the job.
Where did this question come from?
I came to it from two directions. During my early study for jukai, the ceremony in which a Zen student formally receives the precepts, I read Nancy Mujo Baker’s Opening to Oneness. More recently I have been working through Gert Biesta’s Good Education in an Age of Measurement, and his third chapter, on education between accountability and responsibility, kept sending me back to my notes from Baker.
This is a borrowing, and I want to be clear about that. Biesta is writing about the culture of accountability in schools, and in this chapter he leans heavily on the sociologist Zygmunt Bauman. Baker is writing about Zen ethics, and she leans on Dogen, the 13th-century founder of the Soto school. Neither of them is writing about AI policy. Bringing them together is my move, not theirs. Baker, at least, seems open to this kind of borrowing. She has said she does not think of jukai as a rite of conversion, and she would like to see the precepts taught more widely, including by teachers who do not consider themselves Buddhist.
What does Bauman mean by responsibility?
Bauman’s argument rests on a distinction between ethics and morality. In Biesta’s reading, morality concerns the part of human thought, feeling, and action that deals with right and wrong, while ethics refers to rules, codes, and norms. Ethics is the codification of moral action, and it carries an assumption that a moral life is a life of obedience to the code. Bauman describes modernity as the era that believed such a code could be written, one that would be free of ambiguity and leave no dilemma unresolved. What he calls postmodern is the loss of that belief.
Bauman is careful to say this is not the end of morality. It is only the end of codified morality, and it gives no guarantee that people will be more moral without the code. It is a chance and nothing more. The argument for that chance that Biesta finds most convincing is the one I underlined in my copy: following the rules, however scrupulously, will never save us from responsibility. We can always ask ourselves, and others can always ask us, whether following a given set of rules was the right thing to do, and we will never have a conclusive answer. Moral dilemmas are real dilemmas. They are not, in Bauman’s phrase, the temporary and rectifiable effects of weakness, ignorance, or blunders that a better code would eventually fix.
Responsibility in this account has two features worth holding onto. The first is that it is one-sided. Following Emmanuel Levinas, Bauman argues that a moral relationship is not a contract. I am not responsible for another person because they are, or will be, responsible for me. The second is that responsibility is not something I decide to take on. It is already there before I choose anything, which is what Bauman means when he says that I am moral before I think. I can refuse it or forget it, and most of the time the routines of everyday life let me do that without much trouble. I cannot make it go away.
Bauman also explains why codes can feel like an escape. A code treats the moral “I” as one instance of an ethical “we,” in which any “I” can be swapped for any “she” or “he” who follows the same rule. That is the teacher in the hypothetical. Any teacher following the same guidance with the same tool would have produced the same result, which is exactly what guidance is designed to do. It is also exactly what it cannot do, since the student who wrote about his grandfather was not writing to any teacher.
What do the precepts add?
When I first read Bauman’s insistence that the “I” cannot be added up into a “we,” I assumed it would put him at odds with Baker. Her book is titled Opening to Oneness, after all. Reading the two together more slowly, I think they are working against the same thing from different sides.
Baker starts with language. The precepts come down from Dogen not as “do not kill” and “do not steal” but as non-killing and non-stealing, and her book is largely about how a person gets from the first form to the second. The “non” form describes a precept that is expressed naturally, without effort and without a sense of separation, rather than a command held at arm’s length. She describes each precept as working on three levels: a literal level, such as not killing even a bug; a relative level that takes moral ambiguity into account, as when the bug is a mosquito carrying malaria; and an ultimate level, the paradoxical level of nonduality, where the precept is expressed from a state of oneness. The first half of the book works mostly at the relative level, and the second half takes up the ultimate level through Dogen.
In my jukai study materials, a collection of essays and talks from several teachers that my own teacher put together, the three levels were sometimes tied to the three treasures. The literal level is dharma, the rule and the way things are. The ultimate level is Buddha, where there is no rule, only oneness. The relative level is sangha, the community, where the rule and the oneness meet in actual situations with actual people. That last pairing has stayed with me, since it puts the hardest ethical work in relationship with others rather than in the rule or in private insight.
The part of her book I return to most is her account of why we fail to live the precepts, and how to work with that failure. It has three steps: 1) extend the narrow, literal meaning of the precept, since non-stealing does not only mean money; 2) know ourselves in relation to the precept, without judgment and without shoulds, simply asking whether I am a stealer, a killer, or a liar; and 3) notice how the precepts begin to appear in action once the first two have done their work. Her question about the second precept stayed with me long after I finished the book: stealing what? Money, time, attention? Her own example is about attention. If someone is answering another person’s question and I keep cutting in because I think I have a better version, I am stealing attention from the one who asked.
She is also clear that reaching a precept naturally is not a state anyone settles into for good. In the next moment, in a different context, she may have to make a conscious choice and practice some restraint. That sounds a great deal like Bauman’s point that dilemmas keep arriving and no code clears them away in advance.
This is also where my own reading of the precepts has always landed. A precept can exist and still, in a given moment, give way. Non-killing does not mean leaving termites to eat through the walls of a house, or letting an invasive insect spread because I could not bring myself to step on it. What I have come to think, reading Baker next to Bauman, is that the rule giving way does not mean the responsibility goes with it. When the precept steps aside, I am still the one deciding, and the killing still asks something of me. That may be the clearest picture I have of what Bauman means when he says the code cannot save us.
What Baker means by oneness is not merging into a crowd. Her concern is with the parts of ourselves we would rather hide or deny, the liar, the killer, and the stealer, and her argument is that welcoming and befriending those parts undoes the opposition between stealing and not stealing. As I read her, hiding them is not only a private matter. If I cannot admit that I am capable of stealing, I am likely to see the stealing only in other people, and that is where my ethics turns into a way of standing apart from them. She also insists the work cannot be done alone. Bringing something out of hiding just for myself is not enough, in her account. It has to happen in the presence of others, which is why each of her chapters closes with partner or group exercises.
This is where I think she and Bauman meet. The “we” Bauman objects to is the “we” of interchangeable rule-followers. Taking a precept only at its literal level makes me one of those rule-followers, since anyone who has not killed anything today has kept it as well as I have. Asking whether I am a killer is a different kind of question in that I can only answer it as myself. Baker is careful, too, that oneness does not erase particularity. Following Dogen, she notes that even in moments of complete absorption things are still what they are, so that when I reach for my coffee I do not pick up the pan. Her teacher Bernie Glassman pictured oneness with others as a right hand putting out a fire on the left hand without first asking whether it should. That image sits very close to Bauman’s claim that I am moral before I think, and it is about as far from a contract as responsibility gets, since the right hand is not waiting to be repaid. It helps me to separate two words here. Otherness is the sense that another person exceeds whatever category I put them in. Othering is pushing someone away so that I do not have to recognize myself in them. Baker’s oneness ends othering. It does not end otherness, and I do not think Bauman would want it to.
What would this look like with an AI policy?
Go back to the teacher with the stack of essays. The question the policy lets her ask is whether she complied, and she did. In the terms of those study materials, the policy lives at the level of dharma, while her real question lives in the sangha, between her and one student. Baker’s steps let her ask that question.
The first step is to extend the literal meaning. The disclosure rule protects trust, and the rule against entering names protects students’ privacy. Both are good rules. But if she asks Baker’s question, stealing what, the answer may be larger than either rule covers, and Baker’s own example about attention fits closer than I expected. A student who writes about a death in his family is offering something, and a response written by no one in particular may take something from him that the policy has no category for, whether that is attention, the experience of being read, or the chance to be answered by a person.
The second step is to know herself in relation to the rule, without shoulds. This is not a call for guilt. It is just looking: what was I doing when I sent that essay through the tool? Saving time I genuinely did not have, probably. Possibly also using the approved process as a kind of shelter, so that whatever came back was the tool’s response and not mine. Both can be true at once, and seeing them is the work.
The third step is the one that cannot be written in advance. Maybe she reads that one essay herself and writes two sentences of her own. Maybe she changes nothing about her process and simply notices more. Baker’s steps do not produce a new rule, which is the point. If they did, we would be back where Bauman started.
The fair objection, and I hear versions of it in nearly every guidance conversation, is that teachers need clear rules, not koans. Teachers are already carrying more than any job description admits, and asking them to examine every action can turn into one more unfunded mandate. Clear rules also protect teachers when a decision is questioned and protect students from real harms, and a policy that only says “use your judgment” can leave a new teacher with nothing to stand on. I think that objection is right, and nothing in Biesta or Baker argues otherwise. Biesta’s chapter is about finding a new connection between accountability and responsibility, not about throwing accountability out. Baker begins with the literal level of each precept. A student at jukai still receives the precepts, word for word, before spending a lifetime learning what they ask. The termite case carries over, too. There will be moments when a district’s guidance does not fit and setting it aside is the right call, but setting it aside does not hand the decision to anyone else.
Who is the policy for?
That leaves me with a question about my own work. When I sit with a district to draft AI guidance, I am writing ethics in Bauman’s sense. I am trying to make a code clear enough that a busy teacher can follow it, and I would not want to stop doing that. What I am less sure of is whether a document can be written so that it guides without promising to settle anything, so that it hands teachers a floor to stand on without implying that the floor is the whole building. I do not know yet what that looks like on the page, or whether it is something a policy can do at all.
Baker describes something about the jukai ceremony that I keep coming back to. For most of the ceremony, the person receiving the precepts is asked to be one with them, and yet it ends by returning to each precept and asking the person to maintain it. Maintaining a precept, as she explains it, means being willing to look again, over and over, at my anger, my blaming, and my lying, and to keep befriending what I find. A student who receives the precepts knows going in that they will break them, probably often and probably in ways they will not notice for years. Receiving them anyway is not a promise to be perfect. It is closer to agreeing to keep asking. I would like AI guidance to work more like that, though I am aware that “keep asking” is a hard thing to put in a district handbook.
If you want more on Biesta, What Is School For? Leading Conversations With Biesta lays out his three domains of purpose, and What Happens When We Value What We Measure? looks at how a metric can quietly take over our judgment.
If this connects to something you have seen in your own district or practice, I would like to hear about it. You can reach me at licht.education@gmail.com, and there are more articles, tools, and resources at bradylicht.com.
