Legal artificial intelligence is not yet competent or trustworthy enough to replace most of what lawyers do. At its current stage of technological development, AI often produces shallow, inadequate, or even inaccurate results called “hallucinations.” It also exhibits blind spots and biases against which lawyers must guard. So, current scholarship on legal AI sensibly focuses on its uses and its risks given its current limitations. I focus here on its ethical implications given its certain improvement.
Legal AI is improving quickly. Whatever its current limitations, AI will soon be able to do much of the technical and knowledge-based work human lawyers now do. We should not hide from that fact. If human lawyers were ever inclined to see technical expertise or ability to apply substantive knowledge as the heart of our professional role and identity, we can no longer do so. Pragmatically, if we are to remain useful, not to mention employable, we will need to offer something that AI cannot.
There are ethical as well as pragmatic reasons for rethinking the role of our profession in the age of AI. If our work is to be not just profitable to us as professionals but also worthy of us and good for human beings, we should locate our professional ethos in worthwhile human work. Here I identify and highlight the value of professional tasks rooted in human capacities. These include ethical persuasion, meaningful advice, and responsibility. This sort of work, by nurturing the capacities on which it calls, can both make us useful to clients and help us grow and thrive.
Introduction
As a teacher of legal ethics, I have overlapping aims and aspirations for my students. Most of those are pragmatic. Future lawyers need to understand their obligations to serve clients competently, diligently, and responsibly. They need to understand the importance of keeping client confidences and communicating with clients. They need to be familiar with numerous other and more specific duties ranging from avoiding conflicts of interest to not commingling funds to not presenting false evidence in a proceeding. They also need to know the potential consequences of failing to meet those obligations to their clients and their ability to continue practicing as attorneys.
But the thoughtful study of legal ethics also appropriately includes concerns about the character and well-being of lawyers themselves. As students enter the profession, they can and should be thinking about what sort of lawyers they want to be, what they will have to offer their clients and their communities, and how they can build for themselves worthwhile and flourishing lives in the law. A good legal career can be part of a good life, not just a source of funding for it. That is what I hope for the future lawyers I teach.
And the professional world in which they will try to build those lives is changing. One of the most fundamental changes is or soon will be the centrality of artificial intelligence in the provision of legal services. So far, AI is not smart or trustworthy enough to displace most of what lawyers do. Current AI systems suffer from blind spots, biases, and other forms of unreliability. These systems also so far seem to produce relatively shallow advice, betraying its origin in linguistic pattern-finding and probabilistic text generation rather than authentic understanding. For now, AI can thus be viewed as merely another tool—albeit one presenting particular promise and particular problems.
But it would be a mistake to hide behind current technological limitations to avoid confronting what likely lies ahead. AI is improving quickly. Soon it will better than we are at much of what we do. In particular, AI is likely to make subject-matter expertise a much less central part of what human lawyers have to offer. Knowledge of the law will always be useful to lawyers, if only to better use our AI systems. But knowing the law and being able to predict legal outcomes may soon be something AI systems can do quite well. I am not a technologist or a futurist, so I will not try to predict the pace or form of advancements in legal AI. But nor will I console myself with the hope that AI will never be able to do this or that knowledge-based task. We should confront the likelihood that if all we have to offer is expertise, we may soon find ourselves with little worth offering.
One aspect of the profession that may for the time being require substantial human work is business development. That, however, is not my focus here, for three related reasons. First, as central as rainmaking may be to the business of law, I am unwilling to make it the center of our conception of professional character and identity. Second, as the market for legal services evolves, even business development, like marketing in virtually every other context, may itself take new and AI-aided forms. Third, and most essentially, to the extent that business development remains work that human lawyers do well, that will be in large part because they have nurtured the relationship centered aspects of professional responsibility that are the focus of this essay.
So, we need to ask what makes for a good lawyer and good professional relationship in an age when expertise is no longer the center of either. And again, as we ask this question we should aspire to a vision of professional excellence that is not just useful to the client but also good and worthwhile to the lawyer. Granted, when AI does much of the knowledge-based work we once did, our most urgent and pragmatic priority may be to find work we can still get paid for. And that may be all that some of us can hope to do—if we can even do that much. But we should aspire to more. We should aspire to ways of lawyering that serve clients well while also being worthy of us and conducive to our growth and thriving.
Here I suggest a partial set of elements that such a vision of lawyering might contain. These include practicing empathetic ethical persuasion, helping clients shape meaningful goals, and sharing the weight of difficult decisions. These responsibilities are rooted in human capacities and thus make for work we can potentially do well. They are also burdens that build strengths. They can nurture the capacities on which they call—empathy, honesty, courage, and dignity. And those capabilities are as useful to those who develop them as they are to those on whose behalf they are employed. In Aristotelian terms these are “eudaimonist virtues”—capacities conducive to the living of good, happy, and admirable lives. We ourselves can thus grow and thrive as individuals, professionals, and citizens by nurturing the very virtues that make us valuable to our clients.
Ethical persuasion
One essential aspect of legal work that human lawyers can do better–or at least differently– than AI is ethical persuasive argument. I say “differently” because what matters is not merely that human lawyers may be more persuasive advocates but that we are potentially capable of a certain sort of ethical human persuasion—persuasion grounded in empathy and ethos as well as information and analysis. AI can persuade, sometimes to our detriment. But human beings can approach persuasion in a way that can contribute to our growth. So, even as AI becomes better at certain forms of persuasion, there will be ethical benefits, for both our clients and ourselves, to retaining a central role for a certain human approach to persuasion.
That approach to persuasion emphasizes understanding, as well as possible, the person one is hoping to persuade. Legal argumentation is not a debate contest in which someone will reward us with a win if we make the best or cleverest arguments. We must actually reach and move people; and that means figuring out what will move them. Persuasion happens only if we find arguments that resonate with particular human beings. We must make or find space in the world views of those we hope to reach. I use the term “empathetic engagement” to describe this approach. But whatever one calls it, it requires that we understand what the world looks like to those we hope to persuade. We must put ourselves in the shoes of others and see things as they do. If we are good at that, we will be able to argue more effectively. And if we do that habitually, we may not only better understand those we hope to persuade; we will also better see our world and ourselves and thus ourselves be better able to thrive.
Empathetic engagement helps us see the world better because it shows us the world—including its human struggles, motives, possibilities, and aspirations—from different and challenging points of view. The pragmatic professional necessity of persuasion, coupled with the empathic engagement necessary to persuade, force us into sometimes jarring but often illuminating alternative perspectives. This, in turn, gives us a fuller view and thus a better understanding of the terrain we must navigate if we hope to thrive. Imagine planning how to cross some murky and treacherous physical terrain, a darkly wooded stretch of hill and swamp. You would try to find high ground from which to map out some path. But any one perspective would be limited. Parts of the terrain would inevitably be distant or clouded or hidden behind hills. No matter how lofty one’s perspective or keen one’s sight, any one person’s picture of the world is incomplete. Empathic engagement is like sending out scouts, scouts who can help us fill out our map by telling us what they see from other angles, and what that terrain looks like to them.
We should do this—strive to see the world as others do—even absent the need to persuade. But it can be difficult and painful. The terrain navigation metaphor fails to capture the extent to which our worldviews are tied up with our efforts to make sense of our own lives. When we map out and make sense of the world, we do so in ways that help us think about how we fit in it. Our worldviews are not just mental maps of neutral facts about the world. They are vehicles through which we come to terms with our own lives, understand our own successes and failures, build connections with others, make meaning, and reassure ourselves that the bad things we observe and experience are not entirely our own fault. Having built some tenuous account of life in terms of which our own lives make some decent sense, it can be unsettling and frightening to get a glimpse of things from very different perspectives. From some perspectives, each of us will appear mistaken, inadequate, culpable, or perhaps merely absurd. That is hard to see, even provisionally and momentarily. We know we might better learn and grow and thrive if we could see the world from more angles and open our world views to challenge from alternative perspectives. But something in our psychic self-preservation says: “Don’t go there.” The demands of persuasion force us to go there; and thus can help us grow.
To see how empathetic engagement can help us see not just the world but also ourselves, it is useful to situate this approach to persuasion in a slightly broader context. Aristotle identified three methods of persuasion: logos, pathos, and ethos. “Logos” here means logic, connecting evidence and arguments in ways that support conclusions. This is and will remain central to legal argumentation, but is not always the most effective way to move human beings in the world. “Pathos” in Aristotelian is often interpreted to mean appeal to emotion, and often is. But the Greek term pathos is broader and connotes not just current emotion but also experience and thus world view. In this sense Aristotelian pathos can be seen as pointing to empathetic engagement. “Ethos” means argument through personal credibility—persuading someone that you can be trusted. And when we combine ethos with pathos, we realize that to figure out what will help someone trust us we must understand not only how they see the world. You must also understand how they see us. And that can provide an illuminating way of seeing ourselves.
Illuminating but potentially discomforting. As hard as it is to have our own understandings of the world unsettled, it can be even more painful to have our self-conceptions seen through. When we put ourselves in the shoes of others and see ourselves as they do, we will not always like what that shows us. For example, we may realize that conduct we thought conveyed quiet confidence is in fact interpreted by others as arrogance. Where we thought we were charming or engaging, we may realize that others find us tiresome. What we thought communicated helpful expertise may in fact be read as off-putting mansplaining. We may often realize that words and actions to which we do not attach any significance are in fact being seen by others as salient and troubling. We may even come to learn that there are things we did not even realize we were saying and doing but that are making us less worthy of trust. No wonder we may often avoid seeing ourselves as clearly as we might. Who wants that much light?
But the demands of persuasion can discipline our tendency to evasion and self-delusion. If we want to be persuasive, we must be trusted. But if we are to make ourselves trusted we must know how we are seen. And that can help us see ourselves more clearly. If so, whether or not we succeed in earning the trust that will help us be more persuasive to the benefit of our clients, we will ourselves have benefited greatly, if not painlessly, by gaining some measure of the self-knowledge that is central to our own growth and thriving.
Meaningful advice
Lawyers are counselors as well as advocates. Just as there are distinctively human ways to be persuasive, there are worthwhile human ways in which we can offer advice—ways that can be both valuable to clients and conducive to our own growth. We can help clients think well about their goals, their aspirations, and what it may mean to their lives to use the law in certain ways. We can do this if we are trusted; and we can do it well if we are trustworthy.
We cannot and should not tell our clients what to want. This principle of client autonomy is well stated in American Bar Association Model Rule of Professional Conduct 1.2(a): “ . . . a lawyer shall abide by a client’s decisions concerning the objectives of representation.” But nor should client autonomy become a cover for professional irresponsibility or ill-informed advice. Clients cannot make good judgments about their aims for a legal representation unless we help them understand what the law can and cannot do, how the law interacts with human lives, and what our training and experience shows will be the likely results of the choices they make.
Nor do only unsophisticated clients benefit from help in framing their aims. Granted, those with little experience with the law, such as tort plaintiffs, may be more obviously in need of guidance. But sophisticated corporate clients also look to their general counsels to be part of a team deliberating about ends as they relate to the law. Family law clients may be stressed, and without thoughtful guidance may make decisions they will later regret. Criminal defendants, whether new to the law or distressingly familiar with it, may reasonably look to their lawyers to help them think about the implications of the decisions they make. In all these contexts and more, thoughtful lawyers help clients think through their aims, while ensuring that the clients retain the autonomy to make the decisions that must remain their own.
But different clients will have very different informational needs and often unspoken concerns. Offering good advice calls for human understanding. Lawyers need to be sensitive to human concerns and confusions. We need to be alert to the nuances and limitations of our clients’ ability to communicate their concerns under stress. We should thus also be capable of bringing to light fears and aims that may at first be unexpressed or inchoate. And even the most well-informed advice is not useful unless effectively communicated and understood as trustworthy, so we should be capable of communicating well and effectively—of making ourselves not just heard and understood but also trusted.
This—the thoughtful provision of advice and guidance rooted in human understanding—is valuable work that insightful human lawyers can do well. It is also work from which we ourselves can learn and grow and thus perhaps better thrive. Our work as counselors can benefit us in the many of the same ways as does our work as persuasive advocates. Like persuasion, the thoughtful offering of advice requires that we see things from the perspective of those to whom we speak, in this case our clients. It thus similarly calls upon and nurtures the empathy that broadens our perspective and enriches our understanding of the world. Like persuasion, the effective provision of advice demands that we build trust, ethos. This effort thus also requires us to ask ourselves honestly how we appear to others—to our clients. And this can help us see ourselves.
Our work as counselors can also help us develop the capacity to find or make meaning—to translate inchoate aims into worthwhile aspirations and confusion into some sort of meaningful coherence. Counseling, although done for the client and with the client’s interests in mind, can thus also serve as a form of vicarious ethical training for the lawyer. When we help clients think through how their legal problems connect to their larger aims and aspirations, we are getting practice at thinking through aims and aspirations. When we understand how our clients’ fears and misunderstandings may be informing their judgment and shaping their aims, we are learning how fear and misunderstanding informs judgment and shapes aims. When we help our clients think about what is possible for them, we are getting valuable practice thinking about human possibility. And when we help them think about what various outcomes will mean to them in their lives, we are training ourselves to think about life and meaning.
But not always. Not every client will call upon our help and understanding in these ways; and not every lawyer will be capable of rising to the call when it comes. Some clients will know just what they want to do and will want from us merely technical expertise. We may soon have less to offer such clients, and they less to offer us. But some clients will value thoughtful partners in deliberation about the potential role and uses of the law in their lives. And to those clients we will remain of value. We will remain of value, that is, if we can nurture the capacities required. Not all lawyers will. But the extent that we can nurture those capacities—the ability to think well about aims, aspirations, fears, misunderstandings, and meaning—we will ourselves better grow and thrive. Because that is the essential work of an ethical life.
Shared responsibility
Through our advocacy and our advice, lawyers help clients deal with difficult problems and make decisions to address them; but we can sometimes do something more. We can help them shoulder the ethical weight of difficult decisions. To be clear, we cannot bear that weight for them. Our job is primarily to inform and empower our clients to make their own decisions. But we can also stand behind the advice we give. I suggest here that we can and should stand with our clients and vindicate their trust by sharing a sense of responsibility for what we help them do and decide. This suggestion frames at least four questions. First, what might it mean for lawyers to share a sense of responsibility? Second, why might clients value lawyers who do so? Third, how might this be difficult in an age of artificial expertise? Finally, how might it be good for us as lawyers to bear a share of ethical responsibility for what we help our clients do?
By responsibility in this context, I do not mean liability or external accountability. I mean rather a sense of felt responsibility for potentially troubling and ethically freighted things we help our clients do. Note that troubling need not mean wrong. Some difficult acts are defensible, even appropriate, even necessary. I am not focused here on the question of whether we should help a client do something but on bearing the ethical weight of whatever we decide to help them do.
In thinking about the sense and extent to which lawyers should be or feel responsible for what they advise clients, one relevant consideration is the principle articulated in MRPC 1.2(b), which provides that “[a] lawyer’s representation of a client, including representation by appointment, does not constitute an endorsement of the client’s political, economic, social or moral views or activities.” This is an odd sort of rule. Unlike most ethical rules, 1.2(b) does not require, preclude, or even regulate any conduct on the part of a lawyer. Rather, it purports to tell other people how they should not interpret a lawyer’s conduct. But even granting that we should not interpret representation as endorsement, the principle of 1.2(b) does not run counter to the form of responsibility-taking I describe here. I am not arguing that we should be seen as endorsing our clients’ views. The responsibility-taking that matters here is not about how we are perceived by others. It is about how we connect to or distance ourselves from our own conduct. The salient element of that connection is not approval or endorsement; it is acknowledged agency and felt responsibility. We need not approve of what we help our clients do. The question is whether we face up to our role in it.
At least some clients will value this form of responsibility-taking. Not all clients will care whether or how we think about the advice we give them. This is a question about the nature of the attorney-client relationship; and, as experienced lawyers know, there is no single model for that relationship. There will be situations where clients simply want the best advice and see the lawyer as nothing more than a source of expertise. And as AI becomes more sophisticated, human lawyers will become less valuable to such clients. But at least some clients will value lawyers who offer responsible participation in practical and ethical deliberation. And those are the clients to whom we may still have much to offer.
Recognizing why responsibility-taking may be difficult in the age of artificial intelligence can help us appreciate why it might matter. When human lawyers are the primary source of legal expertise, we are likely to feel at least some sense of some responsibility for what we advise. In such circumstances, which until recently were nearly universal, it is evident that the advice we offer is an expression of our judgement. It is also clearly (if implicitly) a claim that we can be trusted. And thus our advice carries an implicit plea for reliance and trust. But as AI systems become the source of expertise, it will become less clear whether we are even asking to be trusted, let alone willing to vindicate that trust by standing behind what we say. As we delegate epistemic authority, we may be tempted to wash our hands of ethical responsibility.
For now, we can guard against that temptation by using our independent judgment to confirm or double check what our AI systems advise. So far in the development of legal AI, it remains a key component of the core professional responsibilities of competence and diligence that we do not delegate too much to yet unproven systems. But again, AI is getting better quickly. Soon it will be better than we are at much of what we do. And then we may be wise to trust it. Imagine a doctor rejecting the diagnosis of a demonstrably reliable diagnostic tool because of a hunch that the patient does not really have cancer.
Nor can we avoid the ethical problem by saying we should double check and confirm the reasoning or conclusions of the AI systems we employ. That is certainly what we should do now. But it will soon not be possible. It will not be possible because as AI systems become smarter and more reliable, they will also become more opaque. What this means is that even as we realize that our AI systems offer better advice than we could, we will find ourselves less and less able to understand and explain why those systems have offered the advice they have. We may then be tempted, when asked if our advice can be trusted, to think or say: “Well, the machine says so, and the machine is smart and reliable.”
Note that what would be ethically problematic about such a response is not the reliance. In fact, the response might not be unethical in the sense of wrong, assuming the AI is in fact reliable but opaque. It would not necessarily be malpractice. It would, however, be a missed opportunity. Refusing to shoulder responsibility for the advice we offer would represent a failure to do something that clients might value and from which we might ourselves benefit greatly.
How so? How might it be good for us as lawyers to feel responsible for the advice we offer? A full answer to that question is beyond the scope of this essay and is the subject of other work in progress. But the heart of the matter is that responsibility-taking can be a form of ethical weight training. Bearing the weight of what we do, including what we help others do, can build in us a set of capacities vital to our growth and thriving. These, all related in various ways to self-knowledge, include forms of empathy, honesty, courage, and dignity. By facing the impact of our actions on others, we can develop the empathy that helps us see our world and ourselves more clearly. By acknowledging honestly our role and complicity in what we help our clients do, we can develop the capacity for honest self-reflection as opposed to self-deception. By facing rather than fleeing from the potentially unsettling ethical implications of our work, we can nurture in ourselves the courage to confront what may frighten us..
And by bearing the weight of what we do we can vindicate and cultivate our professional and human dignity. Dignity is an elusive idea, so let me clarify. I do not mean dignity in the way human rights law often uses the term—as referring to how people deserve to be treated by others. My focus here is not on how others treat lawyers or how we deserve to be treated. I am focused on how we treat ourselves. Dignity in this context means a sense of self-worth. How do we view and understand our value and worth as a human professional? And dignity of this sort, if it is to ground real growth and thriving, should itself be grounded in something real. Do we claim or aspire to be somehow better than or at least different from the AI systems that will soon become a better source of legal expertise? If so, we should see and bear ourselves accordingly. And one way we can do that is by doing something that AI cannot. We can bear responsibility. We can help shoulder not just the epistemic burdens of legal work but also the ethical weight. We can look fellow human beings in the eye and let them know they are not alone in their efforts to make sense of the law in their lives. We can say: “Trust me.”
Conclusion
I have focused here on the practice of law—on the work through which I hope my students will succeed and build good lives. What will they have to offer, when the practice of law is less about the provision of expertise? And what will the practice of law offer of value to them? But the same considerations should inform the work and lives of those of us who teach and study the law. We have traditionally prided ourselves on our expertise. We study and learn, and thus can teach, the law. But what might it mean to be a useful and thriving human teacher of law in a world where legal expertise is often better provided by AI?
Above all, it will mean preparing our students for that world. If all we give them is knowledge and technical skills, we are not helping them learn things that will make them useful. Instead, we need to think about what traits and capacities they will need to help them be useful and thriving lawyers. This essay so far has been a modest effort to begin to think about that.
But we can also offer and model for our students, and our profession, some of the same things we hope our students in the profession will eventually offer their clients. We can learn to teach with empathy—to see and understand things as our students see them. We can help our students craft worthwhile goals as they work to define for themselves meaningful and satisfying lives in the law. We can help share responsibility for what we do as a profession and a society, rather than hiding behind our role as academics. And just as I have suggested that lawyers themselves can grow and benefit though doing this human work for clients, we as teachers can ourselves learn and grow in analogous ways as we teach.
Sherman J. Clark is the Kirkland & Ellis Professor of Law at the University of Michigan.