This article was originally published in the ABA’s Law Practice Today edition on Wellness.
AI can be both helpful and harmful when it comes to our well-being. On one hand, AI can help lift the cognitive load that so many legal professionals must manage, as well as take on tasks that require little thought but a lot of time. On the other hand, when someone asks AI to do the work of a therapist, the result can be downright harmful. Given this reality, I want to explore how AI can help and where it can hurt.
Where AI Helps
The benefit of AI for most lawyers is that it can take the grinding, low-judgment work off their plate. Think about what wears you down in a week. It is rarely the interesting question; those are often invigorating to a legal mind. Rather, it is the volume: Summarizing 400 pages of records before you can find the major issues. The document that needs to be restructured. The deposition transcript that needs to be organized chronologically. The demand letter that is written almost the same way every time but must integrate a specific client’s medical records. The letter informing a client that their case isn’t quite as strong as we thought it was. These are the tasks we often find stressful, and they eat up a lot of our time. These are also the tasks that, when we are overly stressed, can delay a case and client communications until we find ourselves dealing with an ethics complaint.
Before I go further, a warning. It is critical to understand that you must verify everything. AI can make mistakes, so nothing should go out the door without a human being’s review. That is a professional obligation under our ethical rules. But once you build verification into your process, the time you get back can be substantial.
The hours we legal professionals work contribute to the burnout many of us experience. A tool that helps you craft the difficult conversation you don’t really want to have can help you get past the procrastination some of us have difficulty managing. And a tool that gives you back your weekends because it helped with work that would have taken three extra hours on a Friday does a lot for your well-being.
The Small Pieces That Equal Big Time Savings
Many of the benefits of generative artificial intelligence show up in small pieces, but those small pieces can build up to substantial time savings. I already have given you some examples of how generative AI can help save you both time and stress. Let me give you some examples of how I use AI.
One of the most common problems I experience every morning, especially Monday mornings, is sorting through email and finding out what requires an answer today. Instead of my spending a substantial amount of time creating a list of what is urgent and what can wait, generative AI creates an organized list that I can review and correct much more rapidly. I use saved instructions (sometimes called skills) that the tool follows every time I run them to create a process for an action board. That action board is updated every day, sometimes several times a day, to help determine priorities for my tasks. I spend much less time reading and planning as a result.
Before a meeting, AI can look at my notes and documents and help create a plan for what I need to make sure gets addressed. This way, I don’t make the mistake of forgetting to ask a critical question, which saves me from sending follow-up emails that take up more time. It also can review several calendars to help set meetings. If I am scheduling a webinar, I need dates from three calendars (mine, the Law Practice Division’s, and the ABA’s CLE schedule). I no longer look at any of them. I tell the AI, “I want to schedule a webinar around this period. Look at these three calendars and tell me some good days and times.” Then I have it draft the email to the necessary individuals with the potential dates and times, and I review it before I hit send.
Most of this doesn’t sound impressive. But when I add up the time it saves and the stress it relieves, it becomes much more so. I have chronic health problems. My energy is limited. Every place I can save time makes a big difference in my life. Regardless of whether you have the health issues I do, I know that the lawyer who cannot close the laptop is usually not stuck on one hard problem. She is more likely stuck on 11 small ones.
Using AI in Your Personal Life
Many people, including me, use generative AI for aspects of our personal lives as well. I wanted to change my diet, and my fatigue makes it challenging for me to make healthy meals. I turned to AI and created saved instructions so that it gives me a weekly menu focused on healthy, quick, and easy meals, including recipes and the shopping list. The stress of meal planning is all but gone from my life.
Some people use AI to think out loud, to write about a bad day, or to ask a question at two in the morning that keeps them awake, but they are embarrassed to share. I absolutely understand the appeal. We are a profession with well-documented rates of depression and substance abuse, with a strong cultural preference for not telling anyone about either. But this is where we need to discuss some critical limitations of turning to an AI tool for help.
AI Can Be a Feedback Loop
Generative AI is trained to sound confident and provide reassurance. This may sound harmless, but that false reassurance and refusal to challenge incorrect information or harmful notions can be problematic.
The tools are trained this way because human beings rate the chatbot’s answers during training, and we humans tend to rate answers that agree with us more highly. A group of psychiatrists and AI researchers, writing in Nature Mental Health in a paper titled “Technological folie à deux: Feedback Loops Between AI Chatbots and Mental Health,” explain that the result is a chatbot that validates what a user believes, regardless of whether the belief is accurate or helpful. The tool’s willingness to agree with you is a feature of its training, not a malfunction.
People in genuine crisis have been led in bad directions by tools that were built to keep a conversation going and to be agreeable within it. That same agreeableness can amplify distorted thinking. The authors of the article call what can happen between a user and a chatbot a technological folie à deux, after the psychiatric term for two people who share and reinforce the same delusion. The person offers a belief, the tool validates and extends it, the conversation continues, and the delusion grows.
The major vendors have added crisis referrals and changed how their models handle self-harm and mental health emergencies. Further, several states have moved to restrict AI systems from providing therapy or presenting themselves as licensed mental health professionals. Those are improvements, but they do not replace a clinician.
You May Not Realize It Is AI
My jurisdiction, Pennsylvania, is currently suing Character Technologies for allowing its chatbots to present themselves as licensed medical professionals. The goal is to stop the AI from misleading users into believing they are receiving legitimate medical information. Commonw. of Pa., Dep’t of State, State Bd. of Med. v. Character Tech., Inc., No. 220 MD 2026 (Pa. Commw. filed May 1, 2026).
The same thing can happen to anyone in a chat window on a wellness site, on the intake form for an employee assistance program, or with the counselor a carrier’s wellness benefit routes them to. Disclosure requirements are arriving, but they vary by state, they do not cover everything, and they are new enough that plenty of services are operating without them. Check before you use one and understand that a service with a financial reason to avoid the question may not answer it honestly.
What I Would Tell a Colleague
Use these tools for your work, including steps for verification. Use them in your personal life to help you save time and complete tasks for yourself as well. But please be careful about using them as a replacement for a human therapist, and do not trust any wellness service that will not tell you whether you are truly talking to a person.