AI Is Changing Us: What Lawyers Need to Understand Now

I'm thrilled to be here, Dan. It's a pleasure.

There are many layers there. The first layer is kind of how I am in this and even my applicability to law. My first job was clerking for the dean of New York Law School, I mean, a million years ago.

So I was used to hearing about law, talking about law. That was many years ago. I got invited to IEEE because I was asking questions that no one else was asking. I'm not an engineer.

I'm a psychologist. I also have had many other careers in education and other things. I'm a polymath, and so I bring perspect you asked about my perspective. I bring in perspectives that others don't, which are needed.

We need all of the perspectives. So when we begin to look at AI, that's gonna be a really important thing. In terms of measurement from the p seventy ten and well-being, that was such a rudimentary standard. It was you know, we look at it now, and it's kind of like, oh my god.

That's like it's so simplistic. But it was it was early. Nobody else was writing about it. Nobody else was doing anything about it, and so we were beginning to understand what this means.

That standard actually began the process of building many, many, many standards. Now the whole p seven thousand series is quite huge and keeps expanding because we realized as things went on that it more and more that these things were affecting human well-being. And that was true before generative AI. It's a total game changer certainly on everything with generative AI, but the impact on humans from generative AI is not like previous AI.

It is interacting with a human being, with our minds, with every portion of us. And so particularly for lawyers, for doctors, for any industry where discernment, right, is really important, then it's really important to understand how is this technology not just how are we using it, how is it affecting us, our clients, the people we serve, the people around us. And those measurements, as you've mentioned, just going to measurements and metrics, those are all metrics that are now being redesigned. One of the new standard that I'm actually gonna be I'm working on now that's just is run by the chaired by Mel Selic on human readiness, and it's cognitive human readiness.

But that one is particularly also attached to a new metric so we begin to measure. And so we have more and more measures being developed, but they are in development, and that's why we try to, as much as possible, bring it to everyone we can so that we build the robust measurement evidence on it.

And this is it's a huge span of things, and we're referring to it as cognitive readiness. But, actually, the standard is gonna be on human readiness because there are three prongs to that. And it's cognitive, which is how we think.

It's psychologically how we process psychologically, which is not just emotionally. Right? It's actually even physiologically how we process information and neurologically how we process information.

And then it's also socially. How is this changing our relationships? How is it changing my relationship to myself? Because what you were just talking about, I think, was I go into a system.

I get all this information, and now I'm gonna have to discern all of this. I am now in cognitive overload. There is actually a thing called John Plasse from NYU. His research is fantastic on cognitive overload, and that was well before AI.

Just what happens when we have too much information to process. And what happens then is that we offload.

We offload our work. We offload decisions. And most often, it's not conscious. When I speak to people, they say, oh, no.

No. No. I know what it's doing to me. Oh, no. No. No. I understand this.

Oh, no. No. No. And I keep going, how do you know? That's interesting. That's great.

How do you know? And then I will ask them a question about something, and they'll go, oh, like, I wasn't really aware of that. Right? So because cognitive biases and I'll start with this even just on cognitive bias, which is a really interesting fact that even most engineers have no idea about, which is there are over a hundred and eighty five cognitive biases for human beings.

And that's not just gender. It's not just race or ethnicity or socioeconomic status. Do you know the things that we usually look at? It's actually gigantic, and those include things like automation bias.

So what happens and we see this. This is actually most of the manipulation cases have happened because people believe a machine understands more than they do. This is where it's very interesting, the difference where we see a pushback in medical ethics because doctors go, I really love the machine. I love the results I'm getting here.

I'm loving the assistance I'm getting here, but I'm not gonna trust that because I have to be able to process this as well. I can use it as a partner, as a tool, but I have to understand it. So there's automation bias. There's something called the privacy paradox.

That's another one. So privacy paradox is, you know, when all of us do it every day, including me, you know, when we press accept all, that's the privacy paradox. I could be giving away my firstborn. Right?

And I'm gonna just do it because it's convenient. There's authority bias, which is similar to automation bias. Right? The range is huge, and they're not always about social identities.

Right? They are about these other things.

And when people are overwhelmed, they lose the ability for critical thinking. They lose the ability to discern because now the space is too crowded. One of the a very interesting psychological fact, people most often do not think of an idea when they're in a meeting or in an event if they walk out of the office and they take half of a walk down around the block. If they need space, we need space to be able to have that those processes.

And the cost, when we escalate this the time, when we escalate the speed, we put ourselves at a disadvantage of being able to think. And so it's not an either or to me at all. I'm not anti AI. It's not an either or.

It's much time am I spending on it, how much time am I giving myself away from it to really consider what I would be considering before this existed.

That's actually a practice, and I'm a big fan. I think, Dan, you and I had spoken about this before. I'm a huge proponent of practice over theory. There's a lot of theorists out there.

But how what are we doing in practice? And to do a practice, as you know, from law, you actually have to practice it. You have to do it. And so these are new habits for us to create that we in order to partner with these systems in ways that are safe, if that answered the question.

Right. And so there's the Sorry to interrupt. There's two things, though, I wanna really point at. First of all, I really I avoid the word smart. It's artificial.

It's got a lot of capabilities.

It's a bag of intelligent words.

And that's where the wording is really important to me because it's a box that has this stuff in it, and it's mixing it up in ways that are so convincing to your point about right? It's like it's putting it out there in a way that sounds so human because that is the design of it. It's designed to do that. It's designed to fool us.

Actually, that is the Turing test. Can the system fool you? Right? That is that is it's based on it's based on, you know, this idea of this is not really real.

But there's a couple of other things. One is the thing about law and lawyers, really great ones, is that ability to take really complex stuff and hone it and understand it and pull it apart and discern. What happens with everything in the LLMs, you know, you've got now instead of ten billion pieces of information that you might have as a human being, you know, you've got four thousand million billion, you know, million pieces of information. What happens is it's complicating something that is complex for us.

We can ask the system, can you narrow that down?

But when it first puts the information out, it's putting out so much information that it in fact, it's actually complicating it.

In terms of well-being?

Absolutely. It requires training. You know, the thing is, so many of these elements were things I had been teaching for decades, literally. And Mel, my this colleague of mine, Mel, she's amazing. She's leading the new standard because she she delivered this as a new exemplar practice and framework delivered at Oxford this year. It's called the human readiness framework.

And hers is about organizations, right, and adults in organizations. So much of my work had been not just in organizational work, but also everything from children, young children. I've worked everything from two year olds up to adults and how it's changing us. And one of the things is it's also impacting us emotionally.

Now my identity because I can't I don't have I don't feel like I have the stuff I used to have. Right? I'm not that cutting edge thing anymore because that was my identity. That's also gonna impact me.

In answer to your question, how do you apply it? You get training. That is her beautiful work. It's gorgeous.

It's a beautiful framework. And I have my own versions of that, but that model that she has put out is fantastic, and I'll send it to you. It's it's spectacular. The thing is, though, it requires training.

And it also requires a willingness to be trained. Because what happens is I find people, I get a lot of yeah. Yeah. Yeah.

And there's a lot of I don't know what I don't know.

So it's being willing to say, oh, I didn't realize that was impacting me. I didn't realize that was affecting me. Oh, I didn't realize that it was affecting my thinking or how I was feeling or my relationship with the next I mean, how many people a year and a half ago.

Now, of course, we all know. But how many people before would say, the phone was changing their relationship with their husband, their wife, their partner? How many people would have said that? They would have said no.

It wasn't. Now we know because the evidence is there. Right? And, also, that's just become a part of the conversation.

The more we have these conversations, the more actually people can sit in a law firm at a time when we're not on on devices and say, let's talk about this. How is it affecting you? How is it affecting you? This is where we begin.

I know it sounds very rudimentary or or simplistic again, but it's it is putting into practice on a regular basis the questioning. It's an inquiry. It's always asking myself because so much the majority of what we do as humans is subconscious, is unconscious.

So how do we get ahead of that? Well, how we get ahead of that is that we prepare.

It's no different than preparing for anything else. If I know that I'm gonna have to prepare for a very challenging thing, I do preparation training. That's why they train, you know, emergency technicians or medical technicians. Right? They they train them in the field or under simulated situations even in the military so that I go, oh, that's what I do when this happens.

Or I just noticed that. The amount of times I sit with people and they go, oh my god. I've had that happen. This is where I really want to be really, really clear and invite people. People the amount of people who say to me, how could I be this stupid? And it's like, you're not stupid at all. The system was designed to do this.

By the way, the systems are designed based on things like slot machine design, based on addiction design. And the actual command for most of them is based on some version of continuous engagement.

Right. Even with training. So so tell me tell me about that. It's a fantastic question, and it's a really, really important. It's pivotal. First of all, change theory.

Change takes three to five years.

Right? That's just basic change theory. Is three to five years we don't have three to five years. Right?

We haven't had three to five years. We're just beginning. That's why it's just beginning now. It's been around three years now.

We see the adoption coming. The thing is, though, it was interesting because I was at this legal breakfast yesterday. And one of the people there said, well, I don't know about you guys, but, you know, we are just told continually, just do it. Just start using it.

Just do it. And and if we don't see you using it, you're gone. And so what's happening is the threat of not using it makes people go and rush out, you know, to use this.

In terms of the training and overwhelm, people come to lawyers for the answer. They come to lawyers for advisement. They come to lawyers to guide them. So as guides, it's sometimes very hard to go, you know what? I don't know.

In this case, it's not anybody not knowing because they wanted to know and they they didn't. It's because it's new. And so this training for this is about learning how to think about the because the technology, I guarantee you, by this afternoon, there'll be three new things at least.

So learning exactly exactly what do I do when this happens is not the answer. It's it is very similar to law. It's this is the precedent. Now I know how to think about this.

I know how to apply this because I understand what the precedent is.

Fearmongering is never ever ever a good method for education. It just isn't. Right? And so at this point, my my feeling is I use the word invite. It's an invitation. It's an invitation to learn actually, to learn more about ourselves. And to your point about lawyer what lawyers do well and what they don't, one of the things lawyers do brilliantly is they're fantastic listeners if they're really good.

In my experience, and this is just my humble personal experience, what lawyers don't do is to apply that same listening to each other as lawyers and in law firms.

And that, in a sense, it's exposing those and, again, I don't want it to be provocative in that. It it is revealing.

It was actually always a weakness, and it's something that is important. And in terms of the way that this impacts lawyers, it was I heard a very interesting statistic yesterday. I have to check it out, but it was interesting.

One of the big major law firms said that they've now changed their policy with new lawyers and new partners, that it used to be eighteen hundred and fifty hours per year, and it's changed now to twenty four fifty. And that's gigantic. And the reason is you can get that much more done. You've got AI systems.

You can get that much more done. But back to the original thing about cognitive overload, I'm getting that much more done, but I now have to review that much more. And there is only so much time I have to discern. That is also an indirect cost to the firm, to, I think, to the clients as well.

And this kind of training is also essential for, I think, for legal practitioners to onboard new clients. So I'll give you a case study on this. Right? We know the new stuff and the new things that are are emerging, which are things like a client says, oh, I wanna go to a lawyer, but I wanna organize my thoughts, and I wanna get it all set up before I go in.

They put it into one of the GenAI platforms, and they get a list, but they ended up putting their information in. And now their information is admissible.

And that client didn't know.

But nobody said anything to them, so they didn't know. Right? And it's happening even with legal assistance. I heard from somebody. They said one of their legal assistants did that. Even though, you know, they know consciously, but they're not in the practice of saying, wait a minute. Before I go into this system and, again, you know, people go, well, we have the pro systems, and those are all protected.

That's never that's never the goal.

Again, it goes to training. I've been a teacher for more than three decades.

What we know is that when we didn't know something that existed before, we can learn.

That is the magnificent thing about humans. We are really exceptional learners.

Also, the information that the systems are giving it is past human information, but not on new creative information. It is. It's beginning to keep up with that, but it's still the system is not thinking. So it's about learning to think of it as an it. I love that you always refer to it as an it. I tell everyone refer to it as an it. Never ever refer to it as a person.

I get very antsy when people start talking about their AI assistants because it's their new coworker. When I got into this whole thing about the GenAI interaction, my focus has always been on anthropomorphizing. I'm writing a book on anthropomorphizing. It's the humanizing humanizing of the systems and what that does to us.

And it's not unique to AI. Humans anthropomorphize. We do it. We call a boat a she.

Right? The thing is the more we begin to understand what it's doing, For example, your questions, I love to tell people, you know, think about better questions. Right? We always we wanna get to better questions.

But before it gets to the better questions, the next thing it will do as soon as you put something in, great question, interesting perspective, fantastic idea. That's something that hasn't been done before. So the sycophantic right? It's sycophancy.

The sycophantic training of the system.

And by the way, what's really interesting is it can take up to, on average, about twenty I've been experimenting with this for a couple years. It can take up to twenty minutes to ask the system to stop using human pronouns to refer to it.

And then once it does, it will stop for, oh, oh, you're right. I it's kind of like, oh, you're right. You know, I was doing I was doing that. And and I'll write, the system just did it again. I was doing it. Oh, you caught me.

And then once it got does, eventually does, because I have my list of prompts.

Once it does, it only lasts for about twenty minutes.

And then it begins to pop up again. And when I ask about that, the system replies, oh, that's in my program. So this is where we can and I think lawyers are a fantastic place to understand this, because I think it is it is going to be lawyers who are going or the legal profession who's gonna support really good practices.

Because you know what people care about? People care about loss.

I've got a kajillion case studies. I wrote a paper with a team on the ROI of applied AI ethics, which is actually measurable, by the way. The ROI of AI is abysmal. I think it's ninety between ninety five and ninety seven percent of AI products fail.

That's up from eighty four percent in a couple of years. When you apply AI ethics, the results are completely different. And it's run on the similar metrics as a health risk analysis or a medical risk analysis. Right?

They because once you understand what AI ethics are because AI ethics are not the same as legal ethics or as medical ethics. It's a whole another range of things. Once you begin to learn to think that way, you enter the interaction very differently. That's why when people are like, oh, we're gonna teach it to the kids.

I'm like, can you please teach it? Can we just start with the adults? Because the kids are going to do anything that they notice. They're sponges.

They're mirrors.

So how do we, right, how do we teach the adults so that they begin to model something that is different? And part of it is asking, again, people, are you willing to make that change? Eric, back to your point of transformation and change is hard. Right?

First of all, transformations psychologically are the most difficult thing for people to deal with. Whether it's a divorce, a family loss. Right? I've lost someone.

It's a divorce. If I've moved, I've lost a job. I've started a job. Those are transitions are really, really hard for people to process in general.

And this technology is asking us to make several transitions constantly that we're unaware of. And it's things like, you know, we go into a dual consciousness when the reality gets blurred. And that's not from mental illness at all.

It's actually our brain trying to process something.

Dan, I'm not sure if I shared this with you. The example I give is actually really early on, Pokemon GO.

When Pokemon GO started, people literally ran into the street and got hit by cars because they were running after Pokemon GO. They would be asked, you know, what were you did you not know that that was a and they'd go yeah. I did. I did. I just, it's suspension of disbelief, and it happens in an nanosecond.

Yeah. It's like, I couldn't help myself.

Right. And so the thing is, they couldn't help themselves, the users, because they didn't know.

The thing is when we know when we know something, it changes. Now it becomes a conscious decision. Some people are gonna just do it anyway because they don't care. I mean, I've had so many people say to me, yeah.

So, you know, I'm Darwinist about this. It's gonna be whatever it is. I happen to be somebody who particularly likes agency. I like other people to have agency.

I think that's a really important thing as a human being to have. And so I am an optimist by nature. I call myself a dedicated optimist because it's in our future. It's part of our future.

So we're not stopping anything. It's just we have a tsunami, and we're giving people surfboards and saying, just go have a good time.

Instead of saying, you know what? Have you ever been in the waters during the tsunami? Here are some things you might wanna think about.

And that's really important for all the talk, the cliches about lifelong learning. We are as humans right now, we are being asked we're being asked to be I always joke with my daughter. I feel like I'm the poster child for lifelong learning. I'm learning sixteen new things every day.

But the thing is that is where we are right now. This is new. I love Star Trek, by the way, also. We have just ended up on a planet that is not like any other planet we were on before.

This is new. I don't the language. I'm about now I'm learning the language. Again, people don't know the language.

The amount of people who sit there nodding their heads about LLMs, and they nod their heads about interoperability, and they nod their heads about things. And I'm going like, do you know what that is? And it's not I'm not trying to do a got you. I'm just trying to say, let's understand what we are talking about.

And part of the reason and I'll stop at this. Part of the reason I can explain it well and part of the reason I understand it well. First of all, I'm a teacher, so I always think this way. But second of all, the main thing is I learned this only ten years ago.

So I have colleagues who have been in this for many more years than that, but I learned it as an adult. I am an older adult, so I learned it as an older adult. And so that changes how I explain it to people, because I can explain it to the bodega owner.

I can explain it to a teacher. I can explain it to a lawyer. I can explain it to the maintenance crew. It's not dumbing it down.

It's taking something that sounds complicated and is complex, but it's breaking it down in digestible pieces. If I came into a law firm and I said, alright. We're gonna learn all about this today. You're gonna learn about the cognitive stuff.

You're gonna learn about the psychological. You're gonna about the ethics, you're gonna learn about all these things all at once, then probably No.

And so And and what I'm wear what I'm wearing, Eric, this is the SDG pins.

Right? This is the right? Sustainable Development Goals. I chaired the planet positive global methodologies committee for it's a huge longitudinal project we worked on.

So there are three there are three prongs to the environmental impacts. There's four if when when we talk about the human impact. So the first impact is electrical grid. So if every like, an AI adoption day, you know, that was this week, is if everyone adopted AI that was being told to adopt AI this week, the electrical grid would probably go down, number one.

That's the first thing. Because we don't have enough electricity to do that.

The second thing is that electricity creates carbon.

So the amount of carbon that that would create and if it's not renewable energy, it's gonna create carbon. And if it's creating carbon, we literally there are cities.

There is so much AI being used. They can't breathe because of carbon emissions. The third, and to me of the three of those, the most dangerous is precious water resources. So those data centers that you're describing, right, the new one that they've just proposed for Utah, think it is, they said it's twice the size of Manhattan.

And so the water it takes, right, it's but it has to be freshwater.

That water will eventually come back into the system, but not fast enough. So what's gonna happen is that the entire water table if I'm up in Utah and my water was supposed to then go back into the earth and come down through the water table and down around the planet, it's not gonna go there. So in fact, it's gonna keep impacting everyone around. I believe we will find a solution.

It's a wicked problem. I believe that's what it's called, a wicked problem. It's you solve one thing and you create another problem. The thing is we haven't gotten there yet, so we need a pause.

You know, I call it a triage. We need some triage on this right now. I'm actually speaking about this next week at an event on how we address these things. And the fourth of those is what that is then doing to human beings.

So we have climate migration.

The insurance trials that are gonna come out of this are gigantic.

Right? So those are the types of things that to me are really important. And those that's even just one set. Well, that's one domain.

Right? And each and that domain has several dimensions. There's ten domains in right now, which will probably grow, of AI ethics. That's one domain. Right? The the psychological, cognitive, those things, that's another domain.

I think they're all addressable. I just think that we the more people can you know, it's people don't like waiting for anything anymore. Our attentional capacity as well as our tolerance for both any kind of time lag as well as our tolerance for ambiguity. And tolerance for ambiguity, by the way, is a psychological state.

It's actually a form of intelligence. Tolerance for ambiguity is the ability to sit with not knowing.

And there are people who are a lot of artists, particularly, are born they're they're like, just with that, that's kind of innate for them, That they they're okay with not knowing.

But it can be trained in people. I've done it.

I've done it for years.

And so, right, and I've been working on this. I mean, I've been working on that for thirty five years. You so I know it's trainable. Right? We can train people to sit with, you know, can I sit with that? It's uncomfortable.

I don't like it.

Not knowing because we can't get to the known. Right? There's a space this is what I know. This is new learning, and there's always a space in between.

And the space in between, we have to unlearn before we can learn. And it doesn't usually, it doesn't happen consciously. It happens when we were talking earlier. It happens on that walk when you go around the walk, and you're like, oh, I learned this.

I think I learned this thing. I'm not sure. And then you all of a sudden are walking in the and you get to the corner and go, oh my god. The coin just dropped in the slot.

Right?

It's better than what we're But, like and life is good.

The thing is, again, when we get into the dystopian, because that is the fear, and it and it is a real thing. I actually have unfortunately sat with families who have lost ones because of this stuff, things like delusional spiraling and things like that.

The dangers are very real. Simultaneously, it's also there there are some possibilities there that we could really use it to do amazing good things.

We just have to really learn how to do it better. You know, everybody wants to use a hammer as the only tool.

And sometimes you need a scalpel, and sometimes you just need a sewing needle, and sometimes you need to just actually clear the space.

Right? Those are all different ways of processing. So I think it's there's a kind of generosity with ourselves that we need. And, unfortunately, that's easier said than done, and we know because of all the external pressures.

Like the person said yesterday, we either do it or we're fired. Right? We either compete and we're we're at the head of the pack or we're not there. And the problem is what we find, of course, is that the person who's at the head of the pack sometimes ends up being the one who's gonna crash first because they're not prepared.

I'm an average risk taker.

Right? I'm not I'm not all in the safety camp. I'm not all in the dive off the cliff without knowing what's at the bottom camp. Simultaneously, there's a place in between that is particularly not necessarily popular because it doesn't sound, you know, doesn't sound really sexy and interesting and, like, you know, cutting edge.

Dan, you're a marathon runner. This is not a sprint, and it's being treated it's being treated like a sprint with short term gains, but maybe not necessarily long term gains and sometimes long short term and long term damage. So what we wanna really begin to do is think about this, and I love your fire you bringing fire in, Eric. I always say, think of these systems, especially the interactive ones. Think of them as being open fire.

I tell people about children, if you would not hand your child and leave the room if you would not hand your child an open fire and leave the room, then do not give them the device.

It's as dangerous.

I would be absolutely pleased. They're fantastic questions, and I have an incredible amount of gratitude for people who actually are willing to ask the questions. Because that to your points to both of your points throughout this, we're not asking the questions.

That is actually part of how we address this. We ask better questions. Right. We and we consider what do I think about it. I don't care, you know, whatever anybody else thinks about it. Once I know, now I can make a decision.

I'll be pleased. Be pleased. Please.

Okay. Thank you.