Latest / Divorce Coaches Academy / AI vs. AI in Conflict
Transcript
- Tracy Callahan: Welcome back to Divorce Coaches Academy. I am Tracy. And the other day I was consulting with a potential client and he was describing the communication between him and his wife. He told me that he uses AI to help him respond to her. Now that certainly wasn't the first time I have heard that. More and more people are using AI to help them navigate. difficult communication during divorce and in co-parenting relationships. And there's some very good reasons for doing that. But as we continued talking, I started to get a better picture of what was actually happening. He would receive a message from her, put it into AI, get help crafting a response, and send it back. She would respond He would put that response back into AI and the cycle would start all over again. And I remember thinking as he was describing it, this is just a circular cycle of unending bullshit. Maybe that's not the most sophisticated way to describe it, but I think it's pretty accurate. Because we now have these two people who are clearly struggling to communicate with one another. Except AI has inserted itself right into the middle of the conflict. One person is using AI to figure out what to say. The other person may very well be doing the same thing. And these increasingly polished messages are traveling back and forth without necessarily moving either person any closer to resolving the actual conflict. It's AI versus AI. While While the two human beings involved remain just as stuck. And that conversation has stayed with me because I think it raises an important question for those of us who work with people navigating divorce and co-parenting. What happens when we get really good at improving the communication without really improving the communicator? So before I go any further, let me be clear that this is not going to be an anti-AI conversation. I use AI. I think it is incredibly useful as a tool. And I think there are ways we can use it in our work and teach our clients to use it that are genuinely helpful. I also don't think AI is going anywhere. So debating whether people should or shouldn't use it seems like a fairly pointless exercise at this point. What I'm much more interested in is how people are using it and particularly what happens when we start relying on AI to do some of the things that we actually need to learn how to do ourselves. Because what this potential client was describing wasn't really a technology problem. It was a conflict problem. The technology had simply given him a more sophisticated way to manage the symptoms. And that's where I started thinking about this difference between helping someone communicate better in a particular moment and helping someone actually become better at managing conflict. And those are not the same thing. And I don't think we can afford to confuse the two. So I like to think about this as sort of the bandage can be helpful, but it isn't the treatment. The analogy I keep coming back to. Is this bandage over a wound idea? Because a bandage has a purpose, right? If you have a wound, you may absolutely need to cover it. You need to protect it. You need to stop yourself from making it worse. You may need something that gives the wound a chance, a chance to settle down. But if there is something festering underneath that bandage and all we ever do is keep replacing the bandage, we have an Actually addressed the problem. I think AI can sometimes function that way in conflict. So let's say I receive a hostile message from my former spouse, and I am immediately activated by it. Before AI, maybe I would have responded immediately, defended myself against every accusation, corrected every factual statement I thought was wrong, brought up three things that happened four years ago. And ended the message by telling the other person that they, they are the reason the co-parenting is impossible. AI does something great, right? It can interrupt that cycle. And that's a good thing. If the alternative is firing off a response you wrote while you were furious, then please, please, please, by all means, give yourself some space and use a tool that helps you slow things down. And if AI can help you take six paragraphs of defensiveness and turn them into three sentences that actually addresses the issue, yeah, there's value in that. I can put the message into AI, ask for help drafting a neutral response, send three reasonable sentence sentences instead of three inflammatory paragraphs. And guess what? I can potentially avoid an unnecessary escalation. Again. There's real value there. But as a divorce coach, I'm actually interested in what happened between receiving that message and sending the response. Because that's where the work is. Did the client recognize that they were activated? Did they understand why? Were they able to distinguish between the information in the message that actually required a response and the language that was simply Provocative? Could they identify what they were trying to accomplish before deciding what to say? Could they tolerate an accusation without correcting it? Could they recognize when their desire to respond was really about defending themselves rather than solving the problem in front of them? And perhaps most importantly, did they make a decision about how to communicate? Or did they simply ask AI to make that decision for them? Those questions matter because the goal isn't simply to get through this particular text exchange without a blow-up. Especially when we're talking about co-parenting. These people, these two people, may need to communicate with each other for another five, ten, fifteen years. They're going to be hundreds, maybe three. Thousands of opportunities for conflict during that time. We have to be thinking beyond the message sitting on the phone today. Think about how this actually plays out. Right? So let's take a very ordinary co-parenting example. One parent sends a message that says something along the lines of: Once again, you failed to tell me about the school event until the last minute. This is exactly what you always do. I shouldn't have to find out about these things from the kids. I need you to start communicating like a responsible co parent. Now, there may be one legitimate issue buried in that message. There is a school event, and perhaps information about it wasn't communicated. But that's probably not the part the recipient is going to react to. They're reacting to, once again, they're reacting to this is exactly what you always do. They are reacting to responsible co-parent. And within about 30 seconds, they are mentally compiling evidence of the 12 times the other parent failed to communicate something similar. They're thinking about screenshots. They're thinking about the soccer registration from March. They're thinking, are you kidding me? You didn't tell me about the pediatrician appointment last month. And then we all know where this goes. Now let's bring AI into the equation. The parent copies the message into AI and says, my ex is constantly accusing me of things that are not true. Write a biff response. And AI may produce something perfectly reasonable. Something that sounds like, thanks for letting me know your concern. I will make sure you have the information about the school event going forward. I will share school related information as I receive it. Hmm. That may be a perfectly appropriate response. But I am much more interested in whether the client understands why that response is more effective than the three paragraphs they originally wanted to send. What happened to all of those other things they wanted to say? Why didn't we respond to once again? Why didn't we defend whether they are a responsible co-parent? Why aren't we discussing the pediatrician appointment from last month? Because none of those things solve the problem that is actually in front of us. If the client understands that, yeah, great, we are developing a skill. If they don't, what have they learned? The next time the accusation comes in, they're right back in the exact same place. The technology regulated the communication, but the person hasn't necessarily learned to regulate themselves. And that distinction is incredibly important in divorce coaching. So let's let's talk about why communication is not just about finding the right words. So I think one of the most the things that AI is forcing us to recognize, it's a language model, right? Is that effective communication has never really been about finding the perfect words. We tend to focus on the words because they are the part we can see. We can look at an email and say, that's too aggressive, or take that sentence out, or you don't need to explain all of this. But everything that happened before those words were written is actually much more interesting. When I'm working with someone in conflict, I want to understand what they heard when they read the other person's message. I want to know what meaning they assign to it because. What somebody wrote and what they heard them say are not always the same thing. I want to know what happened internally. Did you feel attacked, dismissed, controlled? Did you feel the need to prove something? Did you start building your case before you even finished reading the message? And then I want to know what you're trying to accomplish. Are we trying to resolve a scheduling issue? Or are we trying to get the other person to finally acknowledge that they have been unreasonable for the last seven years? Because again, those are two very different objectives. And only one of them is likely to be correct. Mm-hmm. Yep. So This is where conflict competency starts to develop. It is not in memorizing the right phrases or learning that we should say, I hear your concern, instead of you're being ridiculous. It's in developing the ability to understand what is happening inside us, inside us internally, while conflict is happening around us. And then making some intentional choices. About how we want to engage. And that's harder. It is much easier to find better words than it is to examine why we desperately wanted to say the original ones. AI can give you the words. It can help you soften them, shorten them, organize them, and sometimes tell you that probably you shouldn't send half of them. What it cannot automatically do is make that internal process happen unless we intentionally use it in a way that supports that process. So there's another piece of this that I think is so important, and that is the information we're giving AI in the first place. AI doesn't know your marriage, it doesn't know your co-parent. Doesn't know what happened six months ago unless you tell it. It doesn't know what happened six years ago. It doesn't know your patterns of communication, the agreements you have made, the things that have already been discussed, the dynamics that exist between you or what you tend to do when you feel threatened. Most importantly, it is receiving the conflict through your description of the conflict. That doesn't mean someone is intentionally being dishonest. Most people aren't. But every one of us tells a story from where we are standing. That is our perspective. So if I tell AI, my narcissistic ex is trying to control me again, here's the message they sent. How should I respond? I have already framed the interaction before AI ever sees the message. If I say my co-parent refuses to respect boundaries, that is different from saying my co-parent asked whether we could switch weekends and I don't want to. Those prompts may be describing the exact same interaction, but they create very different contexts. And this matters. Because one of the most valuable things that happens in good coaching is that another human being can become curious about the client's framing. Not challenge it for the sake of challenging it. Not to tell the client they're wrong, not to take the other person's side, but to get curious. What makes this feel controlling to you? When you say they always do this, tell me what always looks like. What boundary do you believe is being violated? Was there an agreement about this? What do you think they might say is happening here? What are you assuming about their intention? And what would change if that assumption wasn't true? That is not about taking someone out of their experience. It is about helping them develop a more complete view of the conflict. And that matters because the way we define a problem determines the options we can see for solving it. There's another place where I think professionals need to be thoughtful about AI. People in divorce often desperately want validation, and that makes sense. Divorce can leave people questioning their judgment, their decisions, their perceptions, and sometimes their entire understanding of a relationship. There are moments when validation is absolutely appropriate and necessary. But validation without curiosity can also keep people stuck. If every time I describe an interaction, I receive confirmation that my interpretation is reasonable, that my boundary is appropriate, that the other person's communication is problematic, and that my response should protect me from them, I may feel so much better. But feeling validated and developing greater conflict competency are not necessarily the same thing. And I think we need to be careful about this as divorce coaches too, because this isn't only an AI problem. A good divorce coach isn't there simply to agree with everything a client believes about the conflict. Our job is not to become one more person sitting on the client's Side of the table saying, you're right. They are impossible. We are certainly there to support the client while also helping them examine the conflict more effectively. Sometimes that means helping them see that the other person's message was provocative. Sometimes it means helping them recognize that they are responding to something that wasn't actually said. Sometimes it means acknowledging that the other person may never give them the validation or accountability they are waiting for, and continuing to pursue it through every email is keeping them tied in the problem cycle. And sometimes it means asking the client to consider whether something they are doing is contributing to the very dynamic they say they want to change. That's difficult work. Yes, it is. It requires trust. It requires context. It requires professional judgment. And it requires us to know when to validate, when to challenge, when to get curious, and when to simply give someone enough room to figure something out for themselves. This is where divorce coaching becomes even more important. With the rise of AI, right? It makes good divorce coaching more relevant, not less. But perhaps not for the reason people initially assumed. If our value as divorce coaches is primarily that we can help a client write an email, then yeah, we should probably be concerned about AI. AI can write a pretty goddamn good email. If our value is giving clients lists of questions to ask their attorney, yeah, AI can generate those lists of questions in no time. If our value is simply organizing information, summarizing options, or producing language, technology is going to continue getting better at all of those things. But that has never really been the real value of the work in divorce coaching. The value is helping the client understand what is happening underneath the communication so they can begin to develop the capacity to navigate conflict. Differently. When a client brings me a message and says, What do I say to this? I don't necessarily want to immediately answer that question. I may want to ask, what do you think actually needs a response to? I may ask what they felt when they read it, what part they keep going back to, or what they desperately want the other person to understand. I may ask what happens if they don't correct the accusation buried in paragraph three. I may ask whether the response or considering moves them towards the outcome they say they want, or simply gives them the temporary satisfaction of saying what they have been wanting to say. And sometimes we absolutely get to the point of drafting language together. There's nothing wrong with that. Modeling effective communication is part of skill development. But the email isn't the work. The thinking that gets us to the email is the work. And I want that client to eventually be able to go through that process without me. Right? Okay. So what happens in divorce negotiations and co parenting communication, right? They're very similar. So I want you to think about a client who receives a settlement proposal. They put into AI and ask, is this a fair offer? And that sounds like a reasonable question. But fair according to what? And who? What matters to this client? What are their priorities? What are they willing to trade? What does this attorney say about the legal implications? What does their financial professional say about the long-term financial impact? What's the history of their negotiation processes? What has already been offered and rejected? What does the client need in order to feel comfortable making a decision? And what does fair even get to the particular person? Because anyone who has worked in divorce for any length of time knows that fairness is one of the most complicated words we deal with. We talk a lot about fairness in our podcast. Right? Two people can look at exact same proposal and have completely different ideas about whether it is fair. Because fairness isn't simply mathematical. It gets tangled up with expectations, history, sacrifice, fear about the future, and sometimes a very deep need for the outcome to say something about what happened in the marriage. AI can help someone organize information or generate questions. It may help them identify things they haven't considered. And those are useful functions. But negotiation is not simply an exercise in determining whether something is objectively fair and then producing the correct response. There are people involved, human beings, there are interests involved, there are emotions involved, there are legal and financial realities involved, there are trade-offs, there's risk, there's uncertainty, and there is a client who ultimately has to make decisions they can live with. Again, the question isn't simply, what should I say? The deeper question is, is how do I think through this? And again, that is where divorce coaching lives. So this is why I don't think our response as professionals should be: don't use AI. It's not realistic, and I don't think it is particularly helpful. Instead, Maybe one of the emergency emerging skills we need to be developing as divorce coaches is helping our clients understand how to use AI in a way that actually strengthens their conflict competency rather than replacing it. Think about the difference between putting an email into AI and saying, write my response versus saying, help me identify which Parts of this email actually require response. Those are very different uses of the same tool. Instead of asking, how do I tell my acts that they're being unreasonable? Perhaps the client asks say I to help separate the factual issues in the communication from the emotional or accusatory language. Instead of saying, rewrite this so I don't sound angry, maybe the client writes the response themselves and then asks. Identify anything in this response that could unnecessarily escalate the conflict. Don't rewrite it for me. Maybe they ask, What assumptions am I making in my response? Or what information would I need before deciding how to respond? Maybe they prompt it, give me three possible interpretations of this message rather than assuming my first interpretation is correct. Or maybe the prompt isn't about the communication at all. Maybe it's what question should I ask myself before I respond to this? Now we're doing something different. We're using AI as a tool to support thinking rather than outsourcing the thinking. I'm gonna say this again. They are using AI as a tool to support thinking rather than outsourcing the thinking. And I think that distinction is going to become increasingly important. Because if the client simply learns that every time they receive a difficult communication, they should copy it into AI and ask for the right response. We may have improved the communication without increasing the client's capacity one bit. They are still dependent, still dependent on having to be able to communicate, right? They have to interact with this person, manage the conflict. It isn't just the divorce coach anymore, it's the technology. And dependency has never been the goal of good coaching. Right? Because if they are always dependent on AI, they're never going to be able to function. So I think this creates a really interesting opportunity for divorce coaches because we can actually incorporate some of this into our work with clients. If a client tells me they're gonna use AI to re rewrite a co-parenting message, I'm not necessarily gonna tell them to stop. I though might ask them to bring one of those exchanges into session with us. Show me the original message. Show me what you wanted to write. Show me what AI suggested. And now let's look at the differences. What did AI remove? What did it preserve? Why do you think the shorter response might be more effective? Which sentence in your original draft was about resolving the issue? And which sentence was about making sure the other person understood how wrong they were? What was difficult about not sending that sentence? That is a very different conversation than simply saying, here, let me help you write this. Now we are taking something the client is already doing and turning it into a learning opportunity, skills development. Over time, perhaps they draft the response first and then use AI only to review it. Eventually, maybe they don't need the review every single time. Maybe they start receiving these messages and recognizing on their own. Okay. There's one actual question in here. The rest of this does not need my attention. That's progress, that's skill development. And that is what divorce coaching is supposed to do. We aren't trying to make ourselves indispensable. We're trying to help our clients become more capable. Eventually, There isn't going to be a prompt box. Right? This is so important because it is another reason why this matters. You can't outsource these skills because there's limitations to it. Not every difficult conversation happens through text or email. At some point, you may be sitting in a room with this person or in mediation, and your spouse says something that catches you completely off guard. You may be on a phone call with your co-parent when the conversation suddenly goes sideways. Not uncommon. You may be standing at an exchange with your child 10 feet away when the other parent says something that makes every part of you want to react. Your attorney may ask you a difficult question. A decision may need to be made in real time. There isn't always going to be an opportunity to say, wait, wait, wait, hold on. I gotta open Chat GBT or cloud or perplexity and ask what they should do. And that's when we find out whether we have actually developed capacity or whether we have simply developed a workaround. Can I recognize what is happening in me before I react? Can I create a pause, even if it's only a few seconds? Can I get curious about what is actually happening instead of Immediately, assuming I know the other party's intention? Can I identify what matters in this moment and separate that from all of the other things I would really like to address? Can I choose not to engage with something that doesn't move us forward? Can I communicate a boundary without turning it into a threat? Can I stay focused on the decision that needs to be made even when? The other person is inviting me into the same conflict we have had a hundred times before. Those are skills. And like every other skill, they have to be developed and practiced if we expect to have access to them when we actually need them. So I think this brings us back to something that is fundamental to divorce coaching: capacity. The goal is capacity. Our goal isn't to have a client who needs to call their divorce coach every time they receive a difficult email. If that is what is happening indefinitely, then we need to look at whether we are actually coaching or whether we have inadvertently made ourselves responsible for managing the client's conflict for them. And I think we need to apply the same standard to AI. I do not want a client to leave this process simply knowing how to get ChatGBT to write a really good BIF response. I want them to understand why a BIF response works in a particular situation. I want them to recognize when less information is more useful than more information. I want them to understand why defending themselves against every allegation rarely produces the outcome they think it will. I want them to recognize when they are becoming activated and have strategies for managing that activation. I so want them to understand the difference between reacting and responding, but more importantly, to have practiced that difference enough times. That they can actually do it. I want them to walk into mediation capable of participating in difficult conversations rather than simply preparing the right things to say beforehand. I want them to be able to co-parent with someone who may never, never communicate the way they wish that person would communicate. Because that's another reality of this work. Sometimes the other person isn't going to change. Sometimes they're going to continue sending these same infuriating, irritating emails. They're going to continue using the phrases that drive you nuts. They're going to continue interpreting things differently, being difficult in ways you wish they weren't. Conflict competency isn't about becoming so skilled that you somehow make the other person behave differently. It's about engaging differently. It's about developing the capacity to make different choices for yourself, even when the other person doesn't. And yes, I want clients to know how to use the tools available to them. AI can absolutely be one of those tools. But the tool should support the development of the skill. It shouldn't become a substitute for developing the skill. Because otherwise we can end up in this very strange place where two people are sending each other beautifully written, perfectly neutral, emotionally regulated, AI generated messages, while underneath those messages are two human beings who are every bit as stuck in their conflict as they were before. We have improved the appearance of communication. Without necessarily changing the people who are communicating. And eventually that distinction matters. So I don't think the question for divorce coaches is whether our clients should be using AI. They already are. And they're going to continue. I think the much more interesting question is what are they learning while they're using it? Are they learning how to get a better response from AI? Or are they learning how to become better at responding themselves? Because they're not the same things. AI can help a client get through the difficult communication that is sitting in front of them today. And sometimes that is exactly what they need. But our work as ADR divorce coaches has to extend beyond today. We are helping our clients build the capacity to navigate the conversation that happens tomorrow. The negotiation that happens next month, the parenting disagreement six months from now, and the unexpected moment three years from now when the person on the other side says exactly the same thing that used to send them over the edge. And perhaps that is the real opportunity AI is creating for our profession. It is forcing us to get much clearer about divorce coaching and what it. Does. We aren't valuable because we know the right words. And we aren't valuable because we can write a better email or tell clients what to do. If that's how we define our value, technology is going to continue challenging us because technology can do more and more of those things every single day. Our value is in the process of helping another human being slow down enough. To understand what is happening, identify what matters, recognize the patterns they are participating in, evaluate their choices. Mm-hmm. Mm-hmm. Evaluate their choices. And to develop the skills to move through conflict more intentionally. Technology can support that work. It can even make parts of that work easier. And I think we should embrace the ways we can do that. But it can't replace the client's need to develop those capacities for themselves. Again, which brings me right back to that conversation I had with the potential client and that circular cycle of unending bullshit. The problem wasn't that he was using AI, the problem was that nothing was changing. The messages kept coming. The responses kept going back. The conflict kept circulating. And AI had simply become another participant in a dynamic that neither person really knew how to get out of. That's where I think we need to pay attention. Use AI to help you slow down. Use it to help you organize your thoughts. Use it to challenge your assumptions. Use it to identify what actually requires a response. Use it to look at language you have already written and consider whether that language moves you toward the outcome you're trying to achieve. Use the bandage when you need the bandage. Sometimes preventing that wound from getting worse is exactly what needs to happen first. Don't mistake covering the wound for healing it. Because ultimately, our work as divorce coaches isn't about giving clients better words to use. It's about helping them become more capable of navigating conflict when the words are theirs. All right, everyone. I wanna thank you so much for joining us yet again on another episode at Divorce Coaches Academy podcast. I hope you will consider joining us again. Take care.