AI and Trust and Estate Law: The Future Is Here to Stay
“AI and Trust and Estate Law: The Future is Here to Stay,” that’s the subject of today’s ACTEC Trust and Estate Talk.
This is Natalie Perry, ACTEC Fellow from Chicago, Illinois. Artificial intelligence will bring advantages and challenges to the legal field. Today, ACTEC Fellow Professor Gerry Beyer of Lubbock, Texas, will delve into how AI is poised to revolutionize the legal landscape by streamlining processes, enhancing research, and injecting some risk into this intricate field of law. Welcome, Gerry.
Introduction: Exploring AI’s Impact on Trust and Estate Law
Gerry: Well, hello everyone, and I am glad to be here. Now, I think what we should do is start by figuring out what we mean by artificial intelligence. Artificial intelligence refers to computers that use sophisticated algorithms to complete tasks and even learn to enhance performance accuracy. Now, as a background, there are four basic types of AI.
First is reactive AI. This is the most basic and common type of AI that provides predictable outcomes based upon input, meaning it gives the same response to identical situations it really doesn’t learn. You’re very familiar with this. It is when auto-complete puts stuff in when you are typing texts and emails and documents, spam filters for email, Netflix suggestions, and things like that.
Limited Memory or Generative AI
The second type of AI, and the one we’re focusing on today, is limited memory or generative AI that actually will learn as it receives more data. Now, it is trained on a specific set of data. It can receive new data from user-entered data and some of these AIs can access the internet to obtain, or we call, scrape new data. Examples of this include Chat GPT, Bard, Lexis Plus AI, Westlaw Precision AI Assisted Research, and so forth.
Other Forms of AI
Now, two other types of AI are science fiction: the theory of mind AI that would understand the world and intentions and predict behavior, and self-awareness; the AI of science fiction, which is sentient and would understand life and death. The recent interest in AI, or the limited memory and generative AI, is because of the tremendous amount of data that it can have. I mean, limited memory AI was created back in the 1950s, but it was hard to accumulate large amounts of data. Now, these programs have literally billions of data points.
Now, you should know that ChatGPT originally only had data in its system up to September 2021. But, starting in September of 2023 it can access internet information if you only pay for the enterprise version- which is about an extra $20 a month-BARD can also use internet data; also, if you have data stored on Google services, it can access your email, documents, etc., to enhance its responses.
Using AI in an Estate Planning Practice
Natalie: How can I use AI in my estate planning practice?
Gerry: There are many different ways that AI can be used. First of all, probably the primary one is to increase the speed and efficiency of document drafting, document review, and summarization. So, you could use it to help draft wills, trusts, and all the other estate planning documents- pleadings, briefs, and client communications. And it will provide you with greater accuracy. It’s like a super proofreader, way better than the normal spelling and grammar checks.
It can check to see if you’ve met certain requirements like the tax provisions there that you need or other requirements of state law. It can look for inconsistencies. AIs will also take an increasing role in estate and trust administration. It can help in making fiduciary investment decisions and administration proceedings. It can create pleadings and provide all the notices you need to creditors, beneficiaries, to heirs. It can help create and render accountings. It can organize tasks such as the steps for an entire trust or estate administration from the beginning to the end.
Recommendations and Tips for Estate Professions Using AI
Natalie: What are some warnings about artificial intelligence that we should keep in mind?
Gerry: Many of you are going to be thinking about this and you’re going to think like a person I once asked out for a date. It’s interesting, but quote, Gerry, I’d rather not. Well, you just can’t ignore the AI developments. Think about our Model Rules of Professional Conduct that most states have. Common Eight, rule 1.1 says, a lawyer should keep abreast of the changes in the law and its practice, including the benefits and risks associated with relevant technology. So, you need to be aware of AI and use it appropriately.
Now, a key warning is that the AI does not think, it does not feel, and it cannot evaluate truth or falsehood. Now humans, just like they do with pet animals and maybe a car or other appliances, you kind of give them human characteristics so that it’s easier to deal with. But remember, the AIs that we have now do not have anything like that. The AI is merely a sophisticated guessing machine that provides responses based on correlating billions of data points. Therefore, it is capable of making up responses, or as we call it, hallucinating.
You may have heard of the case of Stephen Schwartz, an attorney who used chat GPT to find case law to support his case and he included this in his briefs to the court. Well, it turned out that the case names were fake, the citations didn’t exist, and the quotes did not exist. It was all made up, fake. And he and his firm were sanctioned for $5,000. I wondered if that was really what would happen. So, I’ve done some tests myself on very simple questions and found lots of hallucinations. They say somewhere between 30% of the old chat GPT, and 20% of the new, are just hallucinations so it’s not very reliable.
Now recently, Lexis and Westlaw have come out with their AI products, and they claim they will not hallucinate. However, I have yet to have access to either of them to test them myself. Now, you can enhance the quality of your responses by using the following tips to formulate prompts that you give to an AI.
- You want to use examples if you can.
- Avoid ambiguous language.
- Using clear language.
- Stay on the topic.
- Don’t ask yes or no questions. Try to be open-ended. And then provide context, the audience, your goals, et cetera. You sort of need to guide the AI.
- And then be sure the tone of the prompts that you ask match the tone of the responses you seek.
A common example is if you tell the AI, I love you, it’ll come back and probably say, I love you too. But if you tell it, I hate you, it’ll come back and be with a nasty response. So, you kind of want to guide it with the tone of what you want, and that will help enhance your response.
Ethical Issues for Estate Planners Using AI
Natalie: What are the ethical issues involved with using AI and what protective steps can we take?
Gerry: Well, the biggest one, first of all, is confidentiality. Because when you type anything into an AI, it now becomes part of its database. And so, it remembers whatever information you have included. So, if you put any confidential client information, it will now be available to the AI when it responds to other people’s inquiries. So, your inquiries need to be very generic, so they don’t reveal any client confidential information.
There are also systems that will be available that will promise they won’t send the information off-site. So, it would be contained just within your computer systems. And if that is the case, then it would be probably allowable to input confidential information, but you’d have to be sure that your data is safe.
Disclose AI to Clients
Now, you may want to tell the client. Clients should know that you use AI. You want to disclose it in your engagement letter. And there are some states, including Florida, that are now considering a rule requiring clients to give advance permission before AI may be used.
Now, of course, there’s going to be some concerns with that type of rule, because what type of AI is it? It would be reactive AI that you’d have to just ask it to use an AI that can type in suggestions when you type an email message or a Word document? So they’re going to have to define exactly what type of AI they mean. Also, it would be kind of out of context. We don’t require clients to consent that we’re using other legal tools. We don’t ask them, if we’re going to use Lexis or Westlaw, or we’re going to use a form building software. You don’t ask for client consent for that. But maybe some people will think AI is significantly different.
Supervise Staff Using AI
You also need to supervise your staff. All levels, from secretaries to associates and even the senior partners, need to be sure to use AI properly, and a provision should be included in your employee handbook.
An issue that is tough to resolve is billing. How do you bill for something that the AI creates in seconds that it may have taken an attorney hours to draft? Then advertising, can you brag about your firm’s AI skills, or can you advertise that I have a better AI and it is superior or unique to the AI used by other estate planning firms? These are some very interesting unanswered questions.
Conclusion: Embracing the Future of AI in Trust and Estate Law
Just a few things then to conclude. I want to tell you that there are AI shortcomings now. AIs currently will have more difficulty with interpersonal relationships and moral judgment. So, I don’t think you need to worry about being replaced by an AI, but you should seriously consider using AI in your practice to free you from some of the more mundane tasks and allow you to work on more sophisticated matters, providing you use AI with care. Select products with due diligence. There are approximately 50 legal AI products already on the market. Be sure you verify information, protect your client’s confidence, and properly supervise your staff. If you do that, you can professionally use AI and enhance your practice.
Natalie: Thank you, Jerry. That’s very helpful. We appreciate you educating us more on artificial intelligence.
Also may also be interested in:
Latest ACTEC Trust and Estate Talk Podcasts
Explore the complexities of settling trust & estate disputes, from representing unrepresented parties to recent legal cases.
Legal recommendations for estate planners handling clients with diminishing capacity, including ethical considerations, POAs, trust planning, and more.