In the seventh episode of Speaking of AI, LXT’s Phil Hall chats with Pradnya Desh, Founder and CEO of Advocat AI. From serving as a U.S. Analyst and Diplomat to practicing international law, Pradnya is also a visionary entrepreneur who is passionate about using her unique experience to support accessible and safe AI regulations that don’t dull innovation. She’s an innovator driven by the spirit of service and empowering women in tech. Learn more about Pradnya’s insights into AI regulation, AI accuracy, and her experience at the center of the AI revolution.

Table of contents

Introducing Pradnya Desh, CEO at Advocat

PHIL:

It would be an understatement of mammoth proportions to say that today’s guest has an interesting work history. After completing her Master of Science in Economics at Georgia Tech, she went into 10-plus years working for the US government. Initially as an economic analyst and then later as a frontline representative of the government in trade attache, negotiator, and diplomat roles. In parallel with her government career, she completed a Doctor of Law degree, which set her up for the next phase of her career as a legal practitioner and educator. Fast forward to January 2021, when she founded Advocat AI as a generative legal nAI company, with the lofty goal of making contracting accessible to all, while striving to make legal work joyful. It gives me enormous pleasure to welcome today’s guest, economist, negotiator, diplomat, visionary entrepreneur in legal AI, CEO, and World Cup finalist, Pradnya Desh.

PRADNYA:

Thank you, I’m glad to be here.

PHIL:

It’s great to have you here. So Pradnya, we are, I believe, all products of our experiences. How did your quite unique background and experiences in international diplomacy and international law lead you into generative legal AI and the foundation of Advocat AI?

PRADNYA:

I think it led me, the reason that I have taken the path that I have is because I’m really inspired by solving problems. And when I was a diplomat, I saw the economy on the whole and saw the world as a whole and saw different parts of… actions that I could take and policies that I could be a part of to make a difference in bringing about prosperity. And then after being a diplomat, I switched into a law firm and was pretty surprised by what I saw. I thought that being on the front lines, in working directly with people, I thought that the private sector, unlike the government, really had it all figured out. I thought, so the government is slow, the government is known for being slow, but the private sector on the other hand probably moves at light speed and it’s not that way in the legal industry. I found that the legal industry; it’s very expensive for people and companies to access. It’s very slow and because it’s expensive and slow access is very limited. There’s really a system of scarcity, rather a system of abundance and as a diplomat I wanted to bring about abundance and as a legal practitioner I wanted to do the same. And I saw, really, legal tech  as a way to solve that problem

PHIL:

Great. And I think you’ve opened my eyes to what being a diplomat is in a way that nobody has before. And your references to prosperity and abundance. I hadn’t really thought of that as part of the role of a diplomat. But yeah, that’s quite inspiring. In discussions of AI in general, and generative AI in particular, technological limitations, ethical concerns, and legal frameworks to manage those ethical concerns seem to be inextricably linked. Do you see major ethical or legal barriers in the way of AI reaching its full potential?

PRADNYA:

It’s a tough question really because the technology itself is not good or bad. The technology is. And so it’s really our role as practitioners to make sure that it is used for good. And therein really lies the challenge is that it’s important to do so. And there’s not really a framework for doing so. And one thing I really worry about is a bad actor making either an accidental, making a mistake or intentionally doing something harmful and then a reactive regulatory framework coming about. And that will really stifle the potential benefits that we would see in AI. And so we need to act quickly to figure out the correct regulatory framework so that we don’t have this catastrophic event from happening and then the resulting boomerang regulations.

Is it possible that we’re already in a reactionary phase in terms of regulation?

PHIL:

Do you think it’s possible that we’re already in a reactionary phase in terms of regulation?

PRADNYA:

I think we are. We’re seeing that in the EU. We’re seeing that we’re seeing the US regulations trying to get a handle on it. So we’re certainly seeing that. I think it could be worse. I think that the issue is that the industry is moving so fast and the technology is moving so fast that it at the moment may be outpacing regulations. Now over broad regulations can put a damper on that. And so while we are in that mode already, it’s not too late, yet. I think that we can get a handle around it, but it’ll be important for industry and government to have a say and to come together rather than it just being government dictating what should and should not be done. And I say that as you know is coming from government.

Is reconciling different regulations across countries a difficulty?

PHIL:

Yeah, yeah, absolutely. And when I asked that question, I was specifically thinking of EU where it feels to me like maybe we’re in a reactionary phase already that and the EU’s orientation to many things is more restrictive, perhaps than… maybe restrictive is not the best word, protective. They seem to be very driven by protective, being protective of individual rights. Do you see an issue or a difficulty in reconciling different regulatory perspectives globally? Certainly, if you look at it, there is a contrast, I think, between the US and the EU in how they perceive the role of regulation, the importance of regulation, and the extent to which regulation should apply. But if you throw, say, Russia, China, India into the mix as well, then you get very broad differences, very broad range of possibilities there. Is reconciling that a difficulty?

PRADNYA:

So, I think that question is different for an individual company than it is really for governments as a whole. Because what I’ve seen at least recently for tech companies, at least in the United States, is that while they are not certain as to how the EU regulatory framework will apply to them, they’re afraid of offering their most cutting-edge products in the EU. So, I’ve been seeing European consumers complaining about notices they’ve been getting on various software that they use saying not yet available in the EU.

And that’s the reason why is regulations there’s no technical reason for that. It’s simply that as their lawyers are trying to figure out what risks there are in offering their product to the EU, I think that one effect is that legal systems are selecting necessarily what products their people can and cannot get by how restrictive, you didn’t want to say restrictive but I will say it, how restrictive the regulations are.

And so your question really was though about harmonizing regulatory systems. And so if everybody had the same system as the EU, certainly tech companies would have to figure out how to comply. But at least right now, the reactionary, well, rather the reaction from these companies are, let’s not deal with it now, let’s simply block access. And then maybe we won’t have to deal with it while they figure out the right level.

PHIL:

Yeah, I think that’s a great answer to that question. So I guess it means yes. Yes, it is a difficulty and it would appear that rather than attempting to adapt the product to multiple legislations, they’re simply saying, well, you just can’t have the product. At least right now. Which will have an interesting run-on effect. If people feel that they’re being shortchanged, that will put a lot of pressure on governments.

PRADNYA:

Certainly.

As a former government insider, do you think governments will have the necessary agility and capacity to keep up with the rate of change?

PHIL:

In my experience, governments and government-funded agencies, including DARPA and IARPA in the US, have a great track record of supporting and driving innovation.

Great and very long track record of that. But right now we’re seeing massive acceleration in the rate of progress in the development of generative AI. And it seems that the center of gravity in this is very much within the corporate sector. As a former government insider, do you think governments will have the necessary agility and capacity to keep up with the rate of change?

PRADNYA:

So there’s really two parts of that, is because I think that the government and the agencies that you named did a great job in creating grants and creating funding mechanisms to spur innovation. And that seems like an excellent role of government, in fact. But keeping up with the details of the technological innovations, government just, governments cannot move that quickly. Even on issuing these grants, it takes months to maybe, six months to 12 months to get these grants out.

And by then the technology has so significantly advanced that they are funding whatever their funding is not really what was applied for in the first place. But, so I don’t think that the governments are yet able to do that. But a partnership between industry and governments is the only way to keep up.

Is 2024 the year of AI accuracy?

PHIL:

Oh, that’s a great answer. Thank you.  At this year’s World Economic Forum in Davos, executives from some of the top companies discussed how 2024 will be the year of AI accuracy. I can see this from more than one perspective. Is it wishful thinking? What’s your perspective on it?

PRADNYA:

I don’t think it’s wishful thinking because when I mentioned the fear of a bad actor coming in and taking down the system or causing very different regulations, it’s all because of accuracy and I actually call it safety. Is that we’re in the legal industry, is that we have a legal AI product and in having a legal AI product, I consider my number one concern, my number one ethical concern, but my number one concern at all to be safety for those that use our product.

And AI accuracy is an important part of what leads to that safety. And I think that application companies like mine are working very hard to figure out how to take an LLM which while it has quite a bit of material in it, some of it accurate, some of it not, some of it coming from a source that knows what it’s talking about, some of it coming from a source that has no idea what it’s talking about or can hallucinate, as well as draw on information that is not the best source. And so application companies like mine have been working hard to tackle that problem.  Is that to bring a layer of accuracy to the wild and moving changes that are in LLMs.

Which female historical figures have been particularly inspiring to you?

PHIL:

So March is Women’s History Month. And so I’d like to ask a couple of questions related to this. I’ll start with asking, which female historical figures have been particularly inspiring to you?

PRADNYA:

Let’s see, so I have two very different ones. Is that the first I would say Amelia Earhart because female innovators inspire me enormously and in fact innovators somebody who is attempting to do something that has not done before that should be done inspire me and so that’s my first and then the other is Mother Teresa and the reason why is her spirit of service also inspires me and I feel like innovation with the spirit of service is really what drives me.

What do you think is needed to ensure that the representation of women in technology continues?

PHIL:

Wow, that’s great. What do you think is needed to ensure that the representation of women in technology continues?  Do you think this requires proactive affirmative action or do you believe that this will ultimately take care of itself?

PRADNYA:

Well, at least among the US venture capital scene, it’s not taking care of itself yet. And at the moment, we’re not seeing on the horizon that problem being improved. And I say that because only 2.2% of venture capital, goes to company, in the United States goes to companies that are founded by women. And that is a shocking number to me because it’s not that we have a dearth of female innovators, it’s that those female innovators are getting funded at a much lower rate. And if they’re getting funded at a much lower rate, we’re necessarily choosing, the market rather, is necessarily choosing, what gets built and who leads that charge. And I think that in AI in particular, that’s a problem because having representation among all types of diversity is important in AI because an AI can only look like the humans that are working on it and developing it and adding expertise to it.

PHIL:

What do you see as a solution for that? How are we going to move that agenda forward quickly?

PRADNYA:

Yeah, well… for starters, invest in more female founders that can do so, because I think that investing in female founders means that those founders will hire more women. And so while I think it’s important to bring in, we’re talking gender I know, but bring in all types of voices, because as we invest in diverse voices and diverse founders, we’re able to bring more diverse products to market. And so, I would start with simply, say, talk with your investment dollars that’s my answer.

What’s next for Pradnya Desh?

PHIL:

Okay, that’s great. And you’ve once again shifted my thinking. I was perhaps, I think I probably asked that question with quite a lot of optimism and you’ve shaken my optimism a little bit, but I do hope we can actually collectively do the work to, to achieve that diversity, as you said, not just gender diversity, but overall diversity in order to ensure that we actually have a much stronger product development field.

Economic analyst, negotiator, diplomat, legal practitioner, legal educator, AI entrepeneur. What’s next for Pradnya Desh?

PRADNYA:

Well, I really love what I’m doing right now. So I’m the CEO of Advocat AI. I’m not planning to go anywhere, at least for quite a while, because we’re in such an important moment for the world. I love being at the center of… I would call it a revolution. The AI is bringing about a revolution that’s similar to what we saw in the industrial revolution.

I mean, I consider it that important. The Industrial Revolution changed the world by making manufacturing by hand and manufacturing one item at a time and switched that to large-scale manufacturing in which a human and a machine were working together to bring widespread prosperity to the world.

And we did see that happening. And AI is causing no less of a shift, except that this is for knowledge, work, and expertise. We’re seeing, we’ll again see human and machine work together to bring widespread prosperity.  So I love leading an AI company and I’m excited to see what’s next in AI for the world.

AI optimism and perspective changes

PHIL:

Wow. Well, I wish you well with your continued work with Advocat. That’s a very exciting prospect. And I share your view that this is somewhat parallel, analogous to the Industrial Revolution. So yeah, it’s an exciting time to be in the business. It certainly is. Okay, what’s the burning question that I should have asked you but I didn’t? And what’s the answer to it?

PRADNYA:

Interesting. So, I don’t have an answer to that, but I do want to know this. I’d like to turn that around and ask you a question.

PHIL:

Sure thing.

PRADNYA:

So you do this podcast and I think you talk to some very interesting guests and come upon some very interesting insights. And so I’d like to know an interesting insight that you’ve gained on AI in the last several months in the conversations that you’ve had. Where has your perspective shifted?

PHIL:

Um, I, talking to, um, some remarkable people over an extended period. Um, I think, unlike my question regarding, uh, women in technology where you’ve shaken my optimism, I would say that my optimism has been, enhanced, with regard to AI through the conversations that I’ve had. I think that the people I’ve been talking to are consumed by the power for good. No one is unaware of the risks. No one is unaware of the potential for bad, but I think everyone is very much focused on the power to… change things for the better, change society for the better, and to extract the positive potential out of this technology. I think that’s the biggest single takeaway that I’ve had from the last few months.

PRADNYA:

I’m glad to hear that. That’s my experience as well.

PHIL:

I don’t know, maybe if I took an hour or two to think about it, I might come up with something else.  But yeah, multiple people that I’ve spoken to have taken this angle.

And I have spoken to some very interesting people from very different ends of the AI spectrum.  And it’s great to have your legally oriented perspective on this today. I’ve thoroughly enjoyed this. So Pradnya Desh, thank you so much for participating in this. Thank you so much for your insights. I hope that we’ll have the opportunity to perhaps do this again at some point in the future and compare notes on where we’ve been and where we’re headed next.

PRADNYA:

I hope so as well. Thank you.