Leadership In Law Podcast

S02E59 Complex Landscape of IP and AI with Dr. Don Simmonds

• Marilyn Jenkins • Season 2 • Episode 59

Share episode

Trademark attorney Dr. Don Simmonds takes us on a fascinating journey through the complex intersection of artificial intelligence and intellectual property law. As an early adopter who gained access to OpenAI before ChatGPT became publicly available, Dr. Don brings a uniquely informed perspective on how these technologies are reshaping the legal landscape.

The conversation delves into crucial questions facing legal professionals and business owners alike: Who truly owns AI-generated content? How can companies protect their trademarks in an algorithm-driven world? What guardrails should we establish as AI increasingly permeates creative and legal processes? Don provides a clear, thoughtful analysis of these issues, explaining why current legal frameworks place ownership with the creators of AI systems rather than with the AI itself or those providing prompts.

Perhaps most compelling is Dr. Don's personal innovation story. He shares how he developed Broner, a specialized AI tool that transformed his trademark practice by reducing office action response time from days to minutes while maintaining or exceeding traditional quality. By creating a custom database of relevant case law and implementing careful controls to prevent "hallucinations" (such as citing non-existent legal precedents), Dr. Don demonstrates how AI can enhance rather than replace legal expertise.

The discussion also explores practical aspects of working with various AI platforms, the value of effective prompt engineering, and predictions about how intellectual property's value will increase as AI becomes more pervasive. As open-source models like DeepSeek compete with proprietary systems, Dr. Don sees a future where accessibility to powerful AI tools will ultimately benefit consumers while creating new challenges for IP protection.

Reach Dr. Don here:

https://www.bron-ai.com
https://www.linkedin.com/in/d-s-899164166/

The Law Firm Growth Guild is designed to help you learn and use proven marketing strategies, grow your firm smarter, and scale your law firm predictably.

Visit https://checkout.lawmarketingzone.com to find out more and to join the private community.

My team and I are adding new content weekly so you can be intentional about your growth and development each week.

Join our private community, Law Firm Growth Guild, Your Shortcut to Marketing Mastery and More Clients at
https://checkout.lawmarketingzone.com

Ready to level up your law firm marketing? Book a FREE Discovery Call with Marilyn Here: https://lawmarketingzone.com/bookacall

Leadership In Law Podcast with host, Marilyn Jenkins
Powered by Law Marketing Zone®
https://lawmarketingzone.com
A full-service Digital Marketing Agency helping clients increase Leads, Cases, and Profit by getting their digital marketing right.

Subscribe on your favorite Podcast listening platform!

Like, Share, and Review us!

#leadershipinlawpodcast #leadershipinlaw #lawmarketingzone #marilynjenkins



Speaker 1:

Welcome to the Leadership in Law podcast with host Marilyn Jenkins. Cut through the noise, get actionable insights and inspiring stories delivered straight to your ears your ultimate podcast for navigating the ever-changing world of law firm ownership. In each episode, we dive deep into the critical topics that matter most to you, from unlocking explosive growth to building a thriving team. We connect you with successful firm leaders and industry experts who share their proven strategies and hard-won wisdom. So, whether you're a seasoned leader or just starting your journey as a law firm owner, the Leadership in Law podcast is here to equip you with the knowledge and tools you need to build a successful and fulfilling legal practice.

Speaker 2:

Welcome to another episode of the Leadership in Law podcast. I'm your host, marilyn Jenkins. Please join me in welcoming my guest, dr Don Simmons, to the show today. As a trademark attorney with over a decade of experience specializing in intellectual property law and, more recently, artificial intelligence, don brings a unique perspective to the intersection of law, technology and innovation. His work focuses on helping businesses, from startups to industry leaders, navigate the complex legal landscape of IP and AI. Whether it's protecting AI-generated innovations, addressing trademark challenges in an algorithm-driven world, or exploring the ethical implications of AI and creativity, he's dedicated his career to making these intricate topics accessible and actionable for non-legal audiences. I'm excited to have you here, don, welcome.

Speaker 3:

Thank you, marilyn, I'm very happy to be here.

Speaker 2:

Thanks, that's great. You had mentioned in our pre-interview that you had access to OpenAI. Way before it was chat GPT, you had early access. Tell us a little bit about that. That sounds super exciting, knowing what's come out already.

Speaker 3:

Well, yeah, Just by curiosity, I happen to stumble on AI. They did have chat GPT, but the version they had, I think, was like GPT-2 or something like that. The version that everyone was writing about was that November, well, September, I think of the year that it was writing about was that November well, september I think the year that it was released.

Speaker 1:

Shit check.

Speaker 3:

I mean three, but I was there basically because I had curiosity in an idea of AI and AI, just using AI for law, and so I had basically opened an account for those who were developers are interested in in developing things are so are? He submitted my applications because they had to approve here. Then I got approval and they'd have and basically did not pay for a few months and cause I did? I was so busy. But then DPTA came up and was writing about it and I already had access to it and APR had computer access and he was able to actually use it and of course it's amazing.

Speaker 2:

Fantastic. Yeah, everybody, I think, is dabbling in it a little bit. I know a lot of people that are using it on a daily basis, or at least a version, whether it's Claude or Gemini or ChatGPT and, of course, deepseek now, and it does bring up some questions. You know we're creating content, whether it's articles or, you know, just helping answer questions, but some people are doing creative stuff like videos and audios, images. Who owns the rights to AI-generated content?

Speaker 3:

Currently the question. Right now, stellarust is the creator of the AI.

Speaker 3:

There isn't any clear legislation that says AI would be able to own its own creation, and there's good reason for that. How would we determine the AI knew or has the intent to view something with the actual event, I mean the IP that it creates? It's not a lie, it's not independent, and so I equate it to some degree like a child being able to create their own artwork and all that. The child is alive and it doesn't have the ability to transact as an adult. And that's the quality of interaction that you would need in order to give that child the rights over its arts. And you know, that's why they generally have to wait until a certain age to be able to use their art in the sort of way they're going to. Until then the parents basically run, and I think for AI, it would basically kind of continue in that way, in that the creator is the one that would be.

Speaker 3:

Now, at some point, there might be AI that is independent and you know, decide what was there and how, and so that part, you know when we cross that bridge, then we can make a decision on that.

Speaker 2:

Yeah, so I agree. I consider it a tool that you know it's doing buying this on my prompts, you know I'm telling it what I want and it's creating that stuff. So I see it as a tool, so that what it's creating or giving me is something that I created. So I heard other sides of the argument and I, just at this point in time, I completely agree with you. It's not an independent that could do this on its own. It's not coming up with these ideas and doing it. So, thinking about you work with intellectual property, how is AI disrupting the traditional intellectual property laws?

Speaker 3:

I mean, a lot of people are building, a lot of books are being created, a lot of art is being created. Clearly, that's new there. There's almost pretty much no rivet in the IP field, and I think that AI is also benefiting from other copyright meaning it means trading on IP, writing, all that. And that is the area that where there is a bit of controversy, because if you are free being trained on copyrighted material and the person that created that let's you be rewarded with revenues. That created that let's you be rewarded with revenue. As to how to train the AI, so that is going through the coursework, and in here it's difficult to tell, if you just use the AI, whether it was trained on a specific piece of material, because it's, you know, always answer exactly the same way each time it's used. This is a bit of an area that yeah, it's gray and they're trying to figure out what the password is.

Speaker 2:

Well, and I think, was it just maybe six months ago that Microsoft Word upgraded and added one of the things in the options, a tick box that you would allow your documents to be used to train Microsoft's AI co-pilot?

Speaker 3:

No, I didn't. I didn't know that this is a lot of folks are doing that. I know that Facebook is doing that. Yeah, I guess, yeah.

Speaker 2:

Yeah, I think some you have control over your meta. Maybe not, but I think if I'm writing documents, I use Word to format the books that I write. So do you want your creation, maybe before you publish it, to be available to train AI? My answer is no, so I quickly turn that off, but you know, so it is an interesting time to be in. I absolutely agree. Now, if you're thinking about a small business, a law firm, doing some branding or working on that sort of thing, how do you protect your brand in these days with algorithms and automation and all these kinds of things?

Speaker 2:

When you say how do you protect your brand, the client to somebody, the client who has its trademark brand or Well say yeah, yeah, so I've got a trademark of law marketing zone, so is there some way that I might have that trademark might be endangered because of, I don't know? I mean automations, algorithms, ais or what is it training on? Do we need to watch for that?

Speaker 3:

the quick question here. It's really hard to sort of trace how it's being trained. It's really difficult. And there, how is the AI using the training? If it was sponsored by an inquiry or quirk, or if there's anything I could say, it's like all right, you have a trademark. Is it registered somewhere, like in the US, canada?

Speaker 3:

or whatever country you might use to provide services or goods. That's like your first step. At least do that, and then anything else. That's a little more, especially when it comes to AI. It's harder to determine where an AI is imprinting your trademark. It's like written on the website and you're seeing your trademark there. Okay, there's one issue, but responding to a question and then it uses your trademark. How is it using? Is it just responding to your question or is it? You know, it's probably to brand something and use your trademark in that branding process. That's where you'd have to sort of investigate and see if it's something you should be compensated for.

Speaker 2:

Okay. So I guess, when you're thinking about how we're using ChatGPT or any of the AI, I'm assuming we built a GPT, but it's an agent on any other platform, right? So some people are building agents to manage part of their business or part of their workflow. We created a GPT for immigration attorneys that basically has all the immigration law up to date as well, you know and can answer in various languages, and that works incredibly well. You did the. You built one for yourself, right? But you added a layer of software on top. Can you talk about that a little bit?

Speaker 3:

Sure thing, and we actually. I think you should mention that after DPT referred to a specific database, I assume a this law, so that when it was founded it at least notes the legislation. In my situation I was in trademark law we responded to what we call office actions. This is when the government organization sends you a smart saying oh, you applied for a trademark.

Speaker 3:

We've refused it and this is the reason, and usually it's because of confusion, meaning that it's a a trademark. We refuse that and usually it's because of confusion, meaning there's a registered trademark or your trademark is descriptive of an increase and thus we must refuse that.

Speaker 3:

We then receive this office action and then, as long as the client is through, I would draft a response.

Speaker 3:

It would take usually a day, a day and a half to draft the next as a VIN, as a response, hoping that the would prove, or the examining attorney could prove, the response.

Speaker 3:

However, after many years of doing this, even when I was able to change the templates to make things easier, the time was still at least today it's more, depending on how many issues I needed to cut it down soon, much less because I was losing. Just say, I needed to have more time to do other things and didn't have enough time to complete these office actions which I was just seeing was seeing in quite a few of them, and really it effectively. So what happened was I said all right, I had access to CharacterPTA, I have access to the API, what can I come up with? And it took me about three months to create this layer of algorithms where I basically used CharacterPTA to respond to office action, but it would do it in a way that was formulated and provide a draft that was one that most attorneys would have thought useful meaning that this is what they would normally argue when they were sparring to an officer.

Speaker 3:

Argue that when it was firing to an officer, the idea was create a form. During the information the question that would help the GPT answer the GPT would respond by all referring to Key's law that we created in a database Similar to what you have. The database addresses the issue of making up case law, which, to Pettini, is the honest remark, and we didn't want that at all, and thus when we created our own database, which is regular, the GPP would provide arguments, refer to case law that is legitimate and then send it to us, and it would do this within five minutes. So what would take a day, maybe a day and a half is now up to your basic retirement, depending on how long you took them in the absence, and the use rate alarm could take you 20 minutes, some people don't. So the idea was now we have something that used to take hours. Now it takes only a minute.

Speaker 4:

And it is just as weird as an original. And during.

Speaker 3:

Usually in thought. I feel this better. There are less errors and we test. We've actually assessed my writing before GPT the ones that I submitted before, versus the ones I submitted later. Being GPT is as good, if not better, than what I've gotten registration, so we're really happy with it. It's called Brian Ayer, brian B-R-O-N-D-A-Y-E-R, and we are an applicator in the US Attorneys Attorney, the EU Attorney.

Speaker 2:

Fantastic. We'll make sure that link is in the show notes as well. That's great. So what you've done is in. What I've read and spoken to other attorneys about is they built something that does a good or better job with a lot less time so they can do a better representation with their clients. That's fantastic. And now you've expanded from the EU market into the US and the Canadian market. Now are you having it like recite the case law so you can confirm that, yes, it's correct, or whatever cases it's referring to, because I spoke with some other people that are doing that. So whenever they're running, you know whether it's an opinion or whatever they're having it. Cite that so they can confirm that these are actual, real cases instead of like that. What was it? Last year? The brief came out that had cases that didn't exist.

Speaker 3:

Yeah, and that was her. I mean, everyone was excited about the UPT. I was and I was, you know, using it as much as any attorney would use it, just to see, oh yeah, the hallucinations are. We call them the fake case law that was rampant, caused a lot of problems, especially for lawyers that were using using it courthouse, you know, wherever they saw it. Yeah, we, there was no way we were going to get off 5 central. What it does is references database of these and I at the beginning I would check it from time to time just to make sure it was not making up case law or definitely referencing only the mark, warrant and deed provided, and it didn't. It never went beyond, always reference the year the case law was.

Speaker 2:

Okay, so you're making the guardrails, You're giving it the database to make sure it doesn't. Okay, that's a good way of doing it keeping it within the rails.

Speaker 3:

Yeah and obviously there might be a point where GPT is so great, knows all the keys of the law and you'll never have to worry about it. I don't know when that is. It could be a few years.

Speaker 1:

Who knows?

Speaker 3:

But in Surat I'm always going to make sure that we're referencing our own data arrays and our data bases regularly and the best we can. Great Good point.

Speaker 2:

Excellent. Have you tried or tested out any of the other AI platforms?

Speaker 3:

Yes, so I mean there's quite so many.

Speaker 2:

There are now yes.

Speaker 3:

Now you basically I don't think you can go wrong. Yeah, I've been trying.

Speaker 3:

Here, of course, you see here's a new one by if they use cloud. It's a new one by if they use cloud, cloud use cloud. I've uh, lambda models. I've used perplexity, which is actually using therapy here. Uh, deep sea. I've used the new Alibaba model, I can't remember. So I've used them all and they all have their own style, their own way of responding, and that's interesting. I'm learning which one is better for what and for code creation, some learning process right now I don't know have you using? I mean, I think you talked.

Speaker 2:

You said you 've used a couple five point yeah, we built the obviously the gpt, through chat gpt. I've used claude for writing, jim and I for writing, especially if you're writing about google products or something or getting a getting ideas, writing outlines, that sort of thing and recently I found one minuteai or one minuteaicom I think it's one minuteai. Anyway, I have to have to look. It's interesting in that it references you like a desk. So say, you are a social media person, then you need to have writing and social media posts and that sort of thing. It takes and access all the different AI platforms in a dropdown box for a particular task, so it's the software over it, but it has access to all these different ones and I'm telling you once it updates. I had no idea. Like you said, there's so many tools. I haven't played with the Lama or the Alibaba one, but you know, gemini, claude and ChatGPT and DeepSeek are the ones that I typically spend my time on and so it's been real interesting to play with that. I got the lifetime deal way back I think in December, black Friday or something like that, cyber Monday and it's just been a very interesting tool to be able to see some of the other platforms and compare. You know, I'll use the same prompt on a different one just to see what comes up. And speaking of prompts, obviously you know to get the good results you want to have a good prompt. How do you handle improving your prompts or helping people improve their prompting?

Speaker 3:

What kind of sound is that? And this might be strange, but the next high-level paid position or career is actually for somebody who had an English language at a high, at a degree level, because prompt engineering is really about being able to use your language or speak English or any other language, but the idea is to know the language so well that when you're speaking to the AI responds according to what you would like, like the response that you want. And a lot of people make this error and they think oh, it's so hard in years.

Speaker 3:

I mean it's not giving me the answer I need. I would say if you're going to get better, get better at your English. That's part of it. And being in law, you know that the words mean things in law Like they're not real. They won't change the contract. They won't change so many things if you use the wrong word in the wrong way and all these things that you're using the wrong word in the wrong way in common, all these things that the check, the gpt it is linear but still really better states to have a great language. You're using it and I think I think that was already beneficial to me because I knew well. I knew the value, words and how to use them and so it would be hard for me to say, well, just do this and you'll be fine, because I already had an advantage. The best way to approach it might be through at UPT to help you answer questions by asking it to give you the problem.

Speaker 2:

I started to. That's one thing I've heard about is people saying these are the things I'm looking for. Give me a prompt to ask you that.

Speaker 3:

Yeah. Yeah, it's sort of doing it backwards, but it works.

Speaker 2:

Yeah, now have you found that it's also you get better results, and I found this. But I'm just curious of what other people are doing If you tell it what you want it to act as. So if I'm looking to do something that's marketing or copywriting or you know whatever it happens to be, if I tell it to be an expert at that, in the voice of, say, dan Kennedy or you know thinking of a Zig Ziglar, you know salesperson or marketing, so I'm giving it a voice, then it turns around and gives me the answer based on what I'm looking for more. Do you find that you do the same thing, or is that something that you'd recommend as well?

Speaker 3:

I definitely, especially when it becomes, where it comes to legal matters, and you go, you're an attorney or you're a legal and you're in this. Please give me a response based on this information and it does provide a response than if you just ask the question. The context helps it really does.

Speaker 2:

Yeah, yeah, and it's a completely different answer than if you just ask a question. So that's very interesting. Well, in your opinion, what do you think is the future of IP law when you're looking at as AI continues to disrupt?

Speaker 3:

industries, because I think it's going to be more apparent as things move, because I hear a strain in these AIs, and AI will be the thing I think will be one of the biggest revolution, even after the Asian. For a long time we have been talking about it, many years. A lot was assumed to be here, but we just didn't know where. Now that it's clearly making headway and this is only what. A couple of years after QPT, and we're having a million helping 6 billion sorry, I don't know parameters for each version of them they're using it in different ways, like robots now, and computers are automatically part of it, part of apps. Now you almost can't escape it. It's pretty pervasive. But we're still early. We still can choose that.

Speaker 3:

Ip is quite being needed to train this, to make them better, make them much and or the value of ip is going to go up, and what's going on is that people are using ar to create. Are using AR to create things like you want to write order. There's sort of the AI is feeding off AR situation which kind of one of us might not lead to better answers and better no.

Speaker 2:

no, you look in and and I do. I want to be respectful of your time, but I wanted to ask about DeepSeek. So you've spent a lot of time with OpenAI, with ChatGPT, and then you tested some of the others. How do you feel about, or what are your thoughts about, deepseq being open source? So I mean, if you had the power, you could run it yourself, but you know how do you feel about them as a competition, to say, the expensive, high-end ChatGPT.

Speaker 3:

It's funny because we kind of knew this was coming. Meta was releasing LLMs for you know, for 18 years.

Speaker 4:

And.

Speaker 3:

I think that one. They're a source of open source, but I think the difference of it, do you see, was that it was caught a whole lot less trade compared to other models, which I think that was the big game changer. And yeah, I don't mind using either, but currently because I think they're kind of small, but because it's so much cheaper to use DHC it's in a charted series, to use DeepSeq it's in a charted series because in the room that I'm in, I don't know how much the difference there is between the two. There are differences, are they glaring? I don't care.

Speaker 2:

Now I've tested, like I said, the same prompt on both. Now Jim and I in Clotter are different but between ChatGPT and DeepSeq and got extremely similar results. So I'm not seeing a big difference. But I think it might be a little bit genius to release. You know, deepseek is free and just imagine how much training it's getting across the globe right now because so many people are diving in deep because it is, you know, free and then open source. So very interesting. So this has so very interesting.

Speaker 3:

Well, this has been quite interesting. I'm sorry, go ahead. I thought that the open source will allow others to jump into the market. I think it's just going to allow more players, which is usually a benefit for the consumer.

Speaker 2:

Yeah, True, true. Well, this has been really good information and very interesting the IP and AI and it's a good time to have these conversations. So I just want to thank you so much for being on the show. I know that the listeners might want to reach out to you and, of course, talk to you about Braun. How can they reach out to you and connect with you?

Speaker 3:

Yes, best way is by email. I'll do that in a row. It's info at Brauncom. That's I-N-F-O at B-R-O-N-E-Rcom. And they can send me any inquiry and I'll respond.

Speaker 2:

Fantastic, okay, great. I'll make sure that those are in the show notes so that people can reach out to you. And again, thank you so much for your time. This has been a really great conversation. I really appreciate that.

Speaker 3:

Thank you, Maren. I hope to speak again. Yes, I agree.

Speaker 4:

Thanks for joining me today for this episode. As we wrap up, I'd love for you to do two things. First, subscribe to this podcast so you don't miss an episode, and if you find value here, I'd love it if you would rate it and review it. That really does make a difference in helping other people to discover this podcast. Second, you can connect with me on LinkedIn to keep up with what I'm currently learning and thinking about. And if you're ready to take the next step with a digital strategist to help you grow your law firm, I'd be honored to help you. Just go to lawmarketingzonecom to book a call with me. Stay tuned for our next episode next week. Until then, as always, thanks for listening to Leadership in Law podcast and be sure to subscribe wherever you listen to podcasts so you don't miss the next episode.

Speaker 1:

Thanks for joining us on another episode of the Leadership in Law podcast. Remember you're not alone on this journey. There's a whole community of law firm owners out there facing similar challenges and striving for the same success. Head over to our website at lawmarketingzonecom. From there, connect with other listeners, access valuable resources and stay up to date on the latest episodes. Don't forget to subscribe and leave us a review on your favorite podcast platform. Until next time, keep leading with vision and keep growing your firm.

People on this episode

Head Shot

Marilyn Jenkins

Host