Pivoting to WEB3

AI and Computable Contracts are Transforming Global Business with Vinay Chaudhri and Donna Mitchell

• Donna P. Mitchell • Episode 64

Share episode

 Can AI Understand Legal Contracts Better Than Humans? 

Imagine a world where AI-powered contracts make legal agreements faster, smarter, and error-free. No more loopholes, delays, or misinterpretations—just precision and automation at a global scale. 

In this episode of Pivoting to Web3, Donna Mitchell sit down with Vinay Chaudhri, a leading expert in AI and computable contracts, to uncover how AI is transforming legal systems, business agreements, and even global policies.

🔥 Key Takeaways:
✅ What are computable contracts, and how are they different from smart contracts?
✅ How AI is reducing risk, cutting costs, and improving legal efficiency
✅ Why critical thinking & logic are essential in an AI-driven world 

Visit [mitchelluniversalnetwork.com](https://mitchelluniversalnetwork.com) for more updates.

 #AIContracts #ComputableContracts #SmartContracts #LegalTech #Automation #ArtificialIntelligence #Web3 #DigitalTransformation #PivotingToWeb3 #EmergingTech 

About Vinay Chaudhri: 

I am passionate about Artificial Intelligence, knowledge engineering and their application to a variety of domains including finance, education and law. My journey on this path is characterized by leadership roles, groundbreaking research and successful implementation of innovative technologies.

In my most recent role at JPMorgan Chase, my team leveraged large language models for a banker note classification task with a potential to increase the business by $500M. I undertook a careful technical comparison of graph database systems to help inform firm's graph database strategy.

At Stanford University, I helped launch a highly successful initiative to bring logic education to high schools across America, undertook education research necessary to for an intelligent textbook product, and contributed to formation of computational contracts initiative.

At SRI International, I formed and led a team that undertook groundbreaking work in knowledge engineering creating the largest axiomatization for a biology book that was incorporated into an intelligent textbook. I also led the development of an ontology and query manager for an intelligent assistant that was the precursor to Siri.

Connect with Donna Mitchell:

Podcast - https://www.PivotingToWeb3Podcast.com
Book an Event - https://www.DonnaPMitchell.com
Company - https://www.MitchellUniversalNetwork.com
LinkedIn: https://www.linkedin.com/in/donna-mitchell-a1700619
Instagram Professional: https://www.instagram.com/dpmitch11
Twitter/ X: https://www.twitter.com/dpmitch11
YouTube Channel - http://Web3GamePlan.com

What to learn more: Pivoting To Web3 | Top 100 Jargon Terms

Donna Mitchell [00:00:00]:
Welcome to pivoting to Web3 podcast where experts break down AI, blockchain and tech shaping our future. I'm Donna and today's guest has insights you won't want to miss. Let's dive in. Welcome to Pivoting to Web three. And today we have Vinay Chaudhry. And Vinay has an outstanding background. He's been a lot of great corporate places, but at the end of the day he does a lot of research. So this is going to get just a little bit more technical and he gets into computerized contracts, but I'm not going to do all the talking.

Donna Mitchell [00:00:33]:
Vinay, say hello to your audience and ours and let us know how you got into the area that you're in and what do you see going forward. Vinay?

Vinay Chaudhry [00:00:44]:
Yeah, thanks Donovan, for giving me an opportunity to come and speak with you today. I feel blessed because most of the times I am locked in my office working, trying to do research and this is a great opportunity to get out and speak to people who are doing real stuff out there in the real world. So I wanted to share with you one area of my research called computable contracts. My background is in AI research and AI researchers are always thinking about working on problems to, to make computers do those things that humans can do. And contracts is one area of human knowledge which is fairly complicated. And we can say with fair certainty that computers or even the AI with a lot of its hype does not understand contracts as well as humans do. And computable contracts is that area of AI trying to change that to see how computers have as much understanding of a contract as a human would. So that's the overall high level view.

Vinay Chaudhry [00:02:05]:
Do you have any questions on that?

Donna Mitchell [00:02:06]:
Yeah, I do. Is the computerized contract something similar to a smart contract or is it different?

Vinay Chaudhry [00:02:16]:
Yeah. So let me explain to you different levels of computable contracts.

Donna Mitchell [00:02:23]:
Yeah.

Vinay Chaudhry [00:02:25]:
So the most basic one is in every contract we have some data. For example, we may have names of the parties involved, the start date of the contract, end date of the contract, and if it is a loan agreement, it might have interest rate, it might have the loan term and payment due date, etc. All of these things are pieces of data. And traditionally contract would be a PDF document. And the first level of making these contracts computable, understandable is to pull out the data. Maybe that data now lives in addition to the PDF, it also lives in a database so that you can run queries against it and do more effective search. That's I would say the level zero of a computable contract. Then in the level up is what you might call a smart contract where you want to enter into an agreement where you want certain things to happen and you want certain things to follow certain rules.

Vinay Chaudhry [00:03:49]:
Okay. And you know, it could be something as simple as buying something on Amazon, right. So they, they have some policies based on which, you know, you got free shipping or certain discounts. And, and you can, I mean, I like to think of that as a form of smart contract where the computer is following a set of, set of rules. Then there is next level up, which is higher than smart contracts, which is not only you have this data marked out in the contract, not only you have certain things happening that follow certain rules, but every clause in the contract exists in the form of computer code. It's a rule that the computer can understand. Those kinds of things, they don't exist today. There are people who are trying to take existing contracts and trying to convert them into set of rules, but it's fairly difficult, very, very difficult for the machine to do.

Vinay Chaudhry [00:05:08]:
But there is a fourth level which goes beyond all this and that says that, well, when a lawyer is, or any contract authors author is designing a contract, instead of writing English, they should write computer code. Everything should be in computer code. And then eventually you turn that into English. So instead of going from English into code, you start from code and then you get English from there. So these are four different levels. And I would say smart contract is somewhere in the middle in this, in this whole process. And based on what I've seen out there in industry, you know, none of these four exist out there. I mean, I mean there are some implementations of some versions of it.

Vinay Chaudhry [00:05:58]:
But I think there is a lot of room here for technology to improve.

Donna Mitchell [00:06:04]:
So what is the, what is the benefit of having that fourth level? Can you give us an example? Why? Because it's missing. There is a thought that it would be beneficial to have more of that fourth level in the mainstream.

Vinay Chaudhry [00:06:23]:
Yeah. So there are certain kinds of calculations which are only possible if your contract exists in the form of code. And one of the projects I have been involved in at Stanford University is, they call it insurance portfolio analysis. So let's say you have a bunch of credit cards and you have a personal umbrella policy and you're going to take a trip to Europe. Now you have to figure out what new insurances you should get because your one credit card may cover some piece of it, the other credit card may cover something else. Your umbrella, some of the risks, your air through airline, you may have some insurance. But how do these insurances overlap? What is missing and what is it exactly that you should buy for your trip? And currently there is no way to answer that. And if you want to be on the safe side, you just buy more insurance.

Vinay Chaudhry [00:07:42]:
You go to rent a car at the rental agency and they try to sell you insurance and you say, oh, you know, it's only $5 extra, and just to be sure, let me just buy it. That's what ends up happening. So I think if you, if all these policies actually existed as code, then we could automatically calculate where the gaps are and how much insurance is actually needed, which is currently not possible to do.

Donna Mitchell [00:08:14]:
Okay, so it sounds like there is a good need on this. Is this part of the reasoning pieces of artificial intelligence, more of the cognitive side of artificial intelligence? Is that more the fourth level that you're speaking about? Yes, and that's where we're going towards now versus just prompting, get an answer. Prompting, get an answer. We're now moving into the place and space where it'll really help you think, really do some deep thinking, analysis, reasoning and, and even though it may not have happened yet, it'll take all that into consideration in the parameters. So with that said, I'm curious to know, what is it that you have seen that made you really do a deeper research in that area? You didn't see enough of it, but was there anything that really captured your eye or gave you that aha moment that motivated you? Like, I got to look deeper into this, I got it. I got to really try to move this forward. Was there any one thing that you can think of?

Vinay Chaudhry [00:09:25]:
The other great example is health insurance. So as you know, if you're in United States and you go to and you need some procedure done, or you need some healthcare, you have a healthcare need and you call them up and ask them how much it's gonna cost. You cannot get an answer because they will tell you that, oh, you know, you have to put the claim through and then we have to evaluate and we see exactly what gets done. Only then we can tell you how much it's gonna cost you.

Donna Mitchell [00:09:59]:
That's true.

Vinay Chaudhry [00:10:00]:
Right. So I mean, shouldn't there be something better? And why is it that we can't get an answer? Again, part of it is how the software is actually architected. You have to go across different organizational boundaries to get the correct answer. But part of it is that it doesn't exist in the form of code. It just is some PDF document sitting somewhere and some human with some understanding of the document who has to actually make the decision, which makes it very difficult to get these straight answers before getting any procedure done.

Donna Mitchell [00:10:51]:
When I think about that level of learning and the opportunity that's out there from a global perspective, with all the different cultures and geographic boundaries, is there a way to make that work no matter where you are, or do you need to take into consideration some of the governance or biases? How do we make that work so there's a global impact? Is that a possibility or is that something that's coded for the culture or the geographic region or the government that you're exposed to? I got a lot in that question. I'm sorry.

Vinay Chaudhry [00:11:28]:
Yeah, I mean, I'm a technologist and usually people start worrying about these social issues when technology works, but this technology is so far away from being practical that people don't even worry about it because, oh, you know, this is just theoretical. Right. So.

Donna Mitchell [00:11:53]:
So what have you seen that you didn't like or that you, that you, that you get concerned about? Since you're on a technology side, I'll come out of the social and go back into the technology. What, what about the technology did you find to be either a really big challenge or something that you saw that raised an eyebrow, that even though the technology is taking place, it concerns you?

Vinay Chaudhry [00:12:21]:
Well, I, I wouldn't say concerned, per se. I think it is.

Donna Mitchell [00:12:31]:
Where you go, I don't know if I'm. Where you kind of get that eyebrows like, oh, no, maybe not a concern.

Vinay Chaudhry [00:12:40]:
Well, I think what, what I have seen is that there is a lack of clarity and lack of education in the public eye about what is possible, what is not possible, what kind of method we should use in a given situation. You know, that's, there's a lot of lack of education. And part of it is because of the hype that we see from coming from the industry trying to push certain kind of technologies. There is this narrative that we see that, oh, we have this smart piece of AI, it's going to read all of your things and it's going to answer all of your questions now, but the reality is that it's not accurate. You cannot trust the answers. You may not get exactly the answer you're looking for. And if you get the wrong answers, you may spend more time fixing the consequences of the wrong answer. There is that view, and then there is the other approach where you step back and you design your contracts and insurance systems in a way that they are very clearly codified and they can be precisely analyzed by the computer.

Vinay Chaudhry [00:14:20]:
It's going to cost you a little bit more upfront, but downstream you will get the answers that are correct, you can trust. So, you know, explaining people, the trade offs between these two approaches, it's the challenge that's the difficult part where people want some magic solution which is actually not going to work out eventually anyways, as opposed to doing a little bit of hard work upfront and doing it right, which is going to work out, which is going to cost a little bit more upfront. But in the long, long run, it's going to work out. You know, it's, it's difficult for people to opt for this second approach versus the first.

Donna Mitchell [00:15:16]:
So does it work with like all the personalization that's taking place right now with customer service and all the hyper, you already predict what the customer might want or do. Is there analysis of that one on one person with this type of research and conceptualization and everything that takes place is do they end up working or intersect in any way where they kind of leapfrog or, or move into a different space.

Vinay Chaudhry [00:15:51]:
Personalization is an important new development in AI technology and it is very beneficial. But I think that starts to go a little bit farther away from the topic of smart contracts or computable contracts. Because smart contracts are about following a set of rules, putting set of rules in the contract in a code that computer can automatically apply. Right. So and in those set of rules, the contract drafter or the parties who agree on the rules, they matter. Right. So they jointly design what the rules are and then the computer should faithfully apply those rules. Right.

Donna Mitchell [00:16:46]:
So when you. Go ahead. I'm sorry, Yeah, I mean, I think.

Vinay Chaudhry [00:16:50]:
You can bring in personalization for smart contracts also in the sense that given a smart contract you want to be able to have a personalized way of explaining it to different users. Right. Like uneducated person or a less educated person may want a very high level view of it. Whereas if you're talking to a legal expert, you really want to get down to the level of details of clauses. So I mean you do want personalized capabilities like that, but I don't necessarily view them as smart contract capabilities. They are more like information understanding and information presentation, which is a layer which sits on top of a smart contract.

Donna Mitchell [00:17:42]:
So, so in that layer, for those that are listening, what industries or sectors do you see it moving forward in? Like you mentioned insurance, you mentioned the medical field or health. Are there other areas where you find the value would be very impactful to the business in helping them scale or offering services or Improving efficiencies and systems.

Vinay Chaudhry [00:18:05]:
Yeah, So I mean there is the, there is this organization called International Systems of Derivative Agreements. So when these large banks or businesses, they have huge investment portfolios, they're investing in certain securities and then they create derivative products out of that. And there are very complicated rules around when things can be sold and how they are priced. All of that has to be codified. And this organization called ISDA is looking at computable contracts as a way to streamline their operations. And that's a huge business value, huge business potential.

Donna Mitchell [00:18:56]:
Yeah, it is, yeah.

Vinay Chaudhry [00:18:59]:
And then in general, you can think of any automated system wherever you want to automate something based on a set of rules. It could be, you know, when should I turn on my sprinklers versus, you know, when should I fire my nukes? Right. So I mean there's this whole spectrum of automation between something very mundane to something which could affect the whole world. Right. So, so anything which is higher stake, right. Firing the nukes or landing something on Mars, you know, all those things, they have to be controlled very carefully and they have to program very precisely. And that's where you want to use technology like computable contracts.

Donna Mitchell [00:19:57]:
Yeah, you helped me in this last example. You really helped me go where you, where you are with this. So at the end of the day, once you said anything in space that brought me into the thought of climate and you know, the different variations or atmospheric changes and other data with, let's say climate is where this fourth layer would fall, possibly in your research and what you're bringing in the advancement of technology and more in that life sciences direction where there's so much more opportunity with imagination that we wouldn't be able to do as a human as quickly as artificial intelligence, but it takes that same skill and really extrapolates it even further. Am I thinking correctly? Am I catching on?

Vinay Chaudhry [00:21:01]:
Yeah, yeah. I mean, again, I think you're certainly going beyond, you know, my own range of thinking. But, but yeah, I mean, again, I think anything which has far reaching consequences like climate, there you want to want the control to be fairly precise and you want the system to follow exact set of rules instead of, you know, just guessing or predicting.

Donna Mitchell [00:21:28]:
So I'm really intrigued on this conversation now that we're into it. But, but let's go into something fun. What is it that you're doing or what projects are you exposed to that you want to share with us or tell us about that's going to change the way life happens or some, some trends that are taking place. Is there anything else that you see that needs to be discussed or just brought to the forefront for people to think about and say, oh, that's going on. Let me, let me hear about that.

Vinay Chaudhry [00:21:56]:
Well, so apart from smart contracts, the other things that I'm engaged in are education. So I'm involved in a project to teach high school students to think clearly and think logically, think critically. Because it's, that's a skill and we need it in everyday life as we are reading newspapers or we are buying products or somebody's pitching us something, or somebody's presenting certain arguments, we have to be able to evaluate those arguments. Right. And I believe that, you know, that's something that should be taught from early on. Some of the thinking, logical thinking we can inherit, we are born with it, but some of it, it has to be taught. And currently in our schools, the logic courses, they are not like a central part of the curriculum. And I've been collaborating with a colleague at Stanford to bring that to high schools across America.

Vinay Chaudhry [00:23:12]:
I mean, obviously, you know, Stanford team is doing most of the work, but I've been involved in ideating it and promoting it and strategizing how to kind of get it out there. So that's one project I'm involved in.

Donna Mitchell [00:23:29]:
Can I ask you a question before you go to your next project? So it's not my imagination. I'm not seeing a lot of critical thinking going on in the world right now. This is, it's that I'm really seeing what I think I'm seeing that people aren't really critically thinking. And that skill, does that skill have to be taught? The critical think? Is that what you're saying? Not just logic, critical thinking has to be taught?

Vinay Chaudhry [00:23:52]:
Yes, that's exactly what I'm saying. Another project I'm involved in is improving the college level textbooks through more systematic codification of knowledge in them. And this is a project I've worked on for many years. It makes the textbooks easier to understand, it makes them more engaging, it more clearly captures what the author is trying to say and overall a good thing. Students learn better with that. I'm also engaged in an AI workshop which we are hosting next month. It is focused on how we can systematically codify knowledge into computer programs in a way that we as humans can examine the knowledge, we can check it for correctness and accuracy, and we can verify and we can make claims about what, what the software is going to do. That's more of a, you know, very fundamental AI research.

Donna Mitchell [00:25:17]:
Yeah, but that sounds exciting. Would that Also cover maybe any biases or things that. Whether there's bias automatically in it or not or is that something.

Vinay Chaudhry [00:25:28]:
Yeah, so I think the bias, the, the bias exists in this world. And if you're going to train your system based on data derived from the world, the bias is going to come because the world is biased. The bias exists in the data. You use the data to train this thing. This is going to have bias. Right.

Donna Mitchell [00:25:52]:
So how do you fix that?

Vinay Chaudhry [00:25:53]:
So I personally argue for knowledge that is universally applicable. Right. So, so things like gravity, right. If something is going to fall, it's going to fall for everybody. You know, it's not like a man is going to be affected differently than a female is. Gravity is going to work the same for both of them. So that's sort of what I've been arguing and that we should focus on codifying universal principles. That should be our priority.

Vinay Chaudhry [00:26:30]:
And then we should probably enhance them using exceptions because there are always going to be some exceptional cases where that universal thing doesn't work. You know, maybe you're in space and gravity doesn't work the way you want and exceptions. And then you should also take into account perspectives, different people, different people's perspectives. Because on some other things we just cannot have a universal, universally agreeable perspective.

Donna Mitchell [00:27:01]:
And there's so much that you're doing is really in that knowledge area. I really like to have you back. Thank you for giving me the opportunity and reaching out to me.

Vinay Chaudhry [00:27:09]:
Okay, thanks, Donna. It's been a pleasure talking.

Donna Mitchell [00:27:12]:
Thank you very much.

Vinay Chaudhry [00:27:13]:
Thank you. Bye bye.

Donna Mitchell [00:27:16]:
Thanks for tuning into pivoting to Web3 podcast. If you're a developer, innovator or AI expert with insights to share, or if you're looking to partner, let's connect. Visit Mitchell universal network.com and be part of the conversation. Want more content? Check out my playlist for more episodes. See you next time.