Artwork for podcast Creatives With AI
E42 - The Intersection of Artificial Intelligence, Art, and Law with Robert Plotkin
Episode 428th March 2024 • Creatives With AI • Futurehand Media
00:00:00 01:04:03

Share Episode


In this episode, David Brown interviews Robert Plotkin, an expert in technology, law, and artificial intelligence. They discuss the use of AI in contract review and drafting, as well as its impact on patent law. They also explore the question of whether AI can be considered an inventor and the potential changes in patent law if AI were able to draft its own patents. The conversation delves into the intersection of AI and copyright, as well as the ways in which AI can augment human creativity. They conclude by discussing the impact of AI on writing and business. This conversation explores the value of human writers in the age of AI and the limitations of AI in writing. It emphasizes the importance of human experience and emotion in writing, particularly in topics that require empathy and personal connection. The conversation also discusses the potential future of artisanal creatives and the challenges faced by traditional news organizations. It highlights the need for creatives to adapt to the changing landscape and leverage AI as a writing assistant and brainstorming tool.


  • AI is being used in contract review and drafting, providing a mix of standardization and customization.
  • The question of whether AI can be considered an inventor is complex and raises legal and ethical considerations.
  • AI has the potential to change patent law and the evaluation of inventive skill.
  • The impact of AI on copyright and intellectual property protection is an ongoing topic of discussion.
  • AI can augment human creativity by providing new tools and capabilities. Human writers can provide value by bringing their unique experiences and emotions to their writing.
  • AI tools have limitations in capturing the human element and emotive aspects of writing.
  • There may be a future demand for artisanal creatives who offer personalized and unique content.
  • Creatives need to adapt to the changing landscape and leverage AI as a writing assistant and brainstorming tool.

Links relevant to this episode:

Thanks for listening, and stay curious!



Tools we use and recommend:

Riverside FM - Our remote recording platform

Music Radio Creative - Our voiceover and audio engineering partner

Podcastpage - Podcast website hosting where we got started


00:07 - David Brown (Host)

Hello everybody, welcome to the Creatives WithAI podcast.


I'm your host, David, with a very wonky voice today, so I have to apologise for that before we get started.


And on today's show we're joined by Robert Block, and he's a trailblazer at the intersection of technology, law and artificial intelligence, as well as being an author. So he's got a very distinctive, unique view of this whole topic and I think it's going to be really interesting. But Robert has more than 25 years of experience. He's mastered the art of securing software patents for a diverse range of clients. He's worked everything from startups all the way up to giant companies such as HP and 3M, so he's got a again got a good view of the landscape. He's an educated, sorry, he's an MIT educated computer scientist who not only understands the intricacies of software, but it's also a leading expert in AI related patents. He's the author of the Genie and the Machine, which is a seminal work on AI in the patent system, and he's a named inventor on over 25 patents. And today, hopefully, he'll share his unique perspective on the evolving world of patents in the age of AI. Robert, welcome to the show.

01:18 - Robert Plotkin (Guest)

Thanks so much for having me, David. Really glad to talk to you about everything that's going on in AI.

01:24 - David Brown (Host)

Was that a decent recap of your background, or do you want to add anything?

01:28 - Robert Plotkin (Guest)

Yeah, absolutely, and the only thing I'd add is that I am an avid user of AI and I make use of it in my writing, and we have begun to make use of it in our legal work. So we are living it as well as writing about it and working with clients who develop AI.

01:49 - David Brown (Host)

Interesting. I think the legal profession is ripe for disruption with AI on a lot of levels.

01:58 - Robert Plotkin (Guest)

Yeah, I mean. I'll tell you that the first area that's gotten the most use of AI has been contract review and drafting. You can understand why because contracts have the right mix of being, on the one hand, very standardised Lots of contracts are very much the same as each other and yet they also need customization and personal attention. They need a mix, and AI is a really good tool for doing things like reading a 50 page contract and spotting. Is there a clause missing here that's in 99% of other contracts and pointing out to the lawyer Maybe you want to add that confidentiality clause.


That's something that normally human lawyers would do. That could take hours of time and could still be missed by a human. AI is really good at that and yet then it doesn't have to be relied on to do the fixing of the problem. It may make a suggestion, but it still gives the human the opportunity and the ability to then write that clause or select it from a library of existing clauses and edit it. So AI is a perfect tool for that type of problem to get all of the benefits of AI its speed, its ability to find and match patterns really quickly and reliably, while still relying on humans for the things that they do really well, which is the bespoke sort of customization of individual sentences or paragraphs or strategy to meet the client's needs. So that's a first there, but I'm glad to talk also about how AI is starting to hit my field of patent law.

03:42 - David Brown (Host)

Yeah, patent law is a really interesting one because I've talked a lot about copyright obviously in the past and I have a friend who works for the UK IP office in Washington DC and she came on and I think from some of the stuff that I've read recently, that things have moved on even from when she was on, because what she was saying is that anything that has that's been where AI has been used in, anything that it's not certainly not copyrightable and then it's also not eligible for IP protection. Is that the same now that it was then, or has it moved on?

04:23 - Robert Plotkin (Guest)

Just in the last year, this issue of whether you can obtain copyright or patent protection for works that were created by or with the assistance of AI has gotten significant attention from patent offices and courts around the world, and there's been a mix of answers, so I don't want to get too deep into the weeds, but it's interesting. These issues have come up over time, going back even to when cameras were first created. This raised the question If you take a picture of a natural scene, can you obtain a copyright of it? You didn't create the natural scene and it was argued that you just pressed the button. Where is the human creativity there in pressing that button on an automated machine, namely a camera? And the law grappled with that and essentially came up with the answer that you could obtain a copyright on photographs because the human creativity involved was in selecting and framing the shot and there was some sufficient creativity. So this is a question that's at least 100 years old, if not longer. It keeps coming up again and again. Every time there's a new development in technology that can either reproduce works or aid in the creation of new works, and it's come up on the copyright side.


For things like AI generated images. We've all used tools like Dall-E and Midjourney, Stable Diffusion and, on the patent side, for when something like a drug is designed with the assistance of AI, can that be patented? And even, who should be named as the inventor? And can an AI system be named as an inventor or a co-inventor on a patent? These are fascinating questions. I actually think that the question of whether AI can be an inventor, although I think it hits us deeply as humans, isn't the most relevant or important legal question. I actually tend to think that we should be seeing AI more as a tool that we use, just like all the other tools throughout history. It is more powerful, and the way in which it is different in an important way is that it can engage in an automated process that can take an input from a human.


Let's say a description of an article you want to write. You give that to chatGPT and it can then create something that maybe goes beyond what you envisioned. The output that it writes may surprise you. It may even be something you could not have created entirely on your own. In that sense, you could say that's different from a camera. When you look through the viewfinder, you see the image, you frame it and when you press the button, you know essentially what you're going to get out. AI can take steps beyond that, but it is still fundamentally a tool that humans use and fundamentally, the intellectual property system, including both copyright and patent, are designed to give humans incentives to create new works. That's still true in the age of AI, and AI systems don't respond to those incentives. So I don't think there is a reason to give copyright or patent protection to an AI system. Let me just dive into that for a second, because it doesn't respond to those yet.


It doesn't yet, and there may be a day when it does. And if their, if AI systems truly become autonomous, in the sense that they have motivations and desires and drives, and an AI system could wake up in the morning, so to speak, and say I want to think about inventing today, and if it could have legal rights to ownership, and if it could be motivated to invent, for example, because it knows that if it could get patent protection it will earn money, and that would motivate it to invent something, then we might have a reason to grant the status of inventor to an AI system. But until that happens, there's really no reason to name it as an inventor on a patent or an author on a copyright and there's no benefit to doing so, if that makes sense.

09:01 - David Brown (Host)

This, this, it's, it's excellent. You've literally casually wandered into one of my big questions that I had for you, which is around the concept if AI could draft its own patents, how would that change the landscape of patent law and how would that, I mean, and change the relationship with IP and humans as well? I mean, that would be massive, yeah.

09:25 - Robert Plotkin (Guest)

ie in the machine was back in:


Let's use the example of Thomas Edison when he created the first really commercially successful light bulb and the key aspect of that was he had to find a material that would work as the filament. I mean, we're all moving away from old incandescent light bulbs but you probably know there's a filament in it. You drive electricity through it. There's resistance that causes the filament to heat up and emit light and it's not. It's a tough problem to figure out what's the best filament, because you need something that will give off enough light but that also won't burn up and disintegrate, for example when you put electricity right. So there's a tradeoff there.


He tried out hundreds, some people say maybe thousands of different materials. It took a lot of money, a lot of time. Well, over a year, maybe more. He was famous for the phrase genius is 1% inspiration, 99% perspiration. That reflects the kind of arduous experimentation process he went through to find that filament for the light bulb. He settled on carbonised cotton at the time, very, very arduous. Well, now you can put in a set of requirements for a material that you want to try to invent, give it to an AI system and it can, in a computer, generate simulated versions of that, simulate experiments on them using the laws of physics and evaluate how good each material is. People are using this for drug discovery. Moderna touted the fact that it used this to rapidly, significantly speed up its process of developing its COVID-19 vaccine. Now it can't do the entire job, but it can radically reduce the time and increase the speed of experimentation.


So if you take a human, providing the input describing to the AI system what goal the human wants to achieve, and then let the computer do the perspiring that Edison talked about, you can essentially have a human AI system which together is a much more powerful, effective, productive inventor than the human alone. To me, that raises the question of what is the threshold or hurdle you should need to get over to be entitled to a patent when it essentially becomes easier to invent. The question isn't whether the AI is an inventor or should be named as an inventor. But if we really reached a point, let's just imagine you asked if AI became motivated by incentives or conscious. What if there was truly an AI system where you could give it a description of any problem you wanted to solve, any machine you wanted to create, any material you wanted to design, and it could reliably spit out a design that solved that problem? Then you could argue there's no need for a patent system because anyone could create something that solves any problem we need.


But where we live now and where we may live forever, maybe for 10 years, maybe for 100 years, I don't know is some fuzzy area in the middle where AI is making us more powerful as inventors, but not infinitely Powerful.


And so, in a nutshell, there are mechanisms within patent law for evaluating whether someone's invention that they're trying to attain a patent on required sort of enough inventive skill to merit a patent, and that is what I think is getting ratcheted up by AI and the patent system. Patent offices, courts need to be taking this into account so that we don't grant patents on inventions that essentially anyone could have created by putting in an obvious input like a prompt to check GPT and getting something out. It doesn't mean that nothing's patentable, doesn't mean that everything is patentable. It means that we need to sort of recalibrate the patent system to take into account the ways in which AI has augmented our ability, and that type of fuzzy, messy recalibration is difficult. I don't think it's as sexy a topic, as you know, or should an AI be named as an inventor on a patent? But it's really where the rubber meets the road in the patent system and where the hard work is going to need to be done.

15:22 - David Brown (Host)

No, that's, that's a great point, and it's the same discussion that I think. You mentioned Edison and you also mentioned photography. I mean, I remember when digital cameras came in and then you had Photoshop, and then we had this. You know, there was the ringing of hands over Photoshop as well, because it's like you could do things in Photoshop that you didn't capture originally and you could alter images in a way that you could never do that before, and it was just using the software. And so there was yeah, there was a big discussion in the community about digital photography, even photography, and yada, yada, yada.


And here we are now and we've got Adobe, who, you know, arguably has the largest bit of, you know, imagery and video software in the world, who has tripled down on putting AI into all of its tools. So now you can actually do that. So now you can end up with these hybrid images where I can go and take a very small photo of a beach, but then I can expand that into some sort of you know 10 different fantastical scenes, if I want to, by adding to the rest of it. So then, what part of that is me and what part of that is the AI? And how do you, how do you get around that? So at the minute it's still, you know, the photographers or the artist still owns the copyright on that image, but it, you know, what point does it cease to become that? I don't know. Have you had any chats about that? Is there any discussion about that?

16:51 - Robert Plotkin (Guest)

Absolutely. I mean again, again. I want to put it in historical context. This has come up many times. You mentioned digital photography, photoshop, Absolutely, that's one of them. Another is in music. You know, think about, in the early days of hip hop, with sampling, the same question. You used to get copyright on music that you wrote, you wrote note for note or that you performed with instruments in your voice. Then people came along and just reproduced portions of songs that already existed on tape or digitally and then mix them together to form new songs, and it raised the same question.


So in each type of the, each of these situations, we move from humans doing the low level crafting of a work, you could say bit by bit, piece by piece, note by note, word by word, brush stroke by brush, stroke right it, going across the different realms of creativity to moving up to a higher level, where there's pre existing components. In the case of art, it might be an entire image or a piece of an image. In music it would be a sample and then taking those components and combining them together. And now, in the realm and the age of generative AI, we're moving beyond just manually taking those pieces and putting them together to giving instructions to an AI to take those pieces and put them together, sometimes in unpredictable ways. You can imagine, as we keep going up this, these computer science, we call them layers of abstraction to higher and higher levels of design, and in each case, to me, there is still not just room but great opportunity for creativity at the new level at which you find yourself. We just need to look at that at that level.


In the case of inventing, we're moving up from the, from the layer of what you might call structural design, where to design a mousetrap, you would figure out what's the lever and the spring and the platform that I need to put together to what an engineer and we typically call requirements design. Specify what, what is the problem I want this to solve, and maybe at a high level, what are the types of features such a mousetrap might need, and then describing that to an AI system for it to come up with a solution that satisfies those requirements. And it's the exact same thing with art. Now, when you describe, you write a prompt for a tool like Midjourney. There's still great room for creativity. I posted, I reposted on LinkedIn recently I found someone who is a graphic designer who seems to be. I apologise, I forget his name, but he posted a bunch of tokens essentially building blocks for prompts that he has found to be very useful. In Midjourney from a great.


He said he's created 75,000 images using my journey and he has found, through all of that experimentation, very much like Edison that we talked about. He's found different types of tokens that are very useful for use in prompts to provide mid journey to get different types of results, and he posted images he's found and when I looked at those images in the picture paints a thousand words I said wow, those are amazing. I know I could not create those images just by providing Midjourney with my very novice type. There still is skill that is both required and useful at the prompting level, that high level requirements level, and I suspect this is my, this is my hypothesis as AI keeps getting more and more powerful, will keep finding new, higher levels at which to be. I don't think it will ever stop, but maybe I'm wrong.

20:52 - David Brown (Host)

No, I don't either, and it's interesting that you said that. I think spoiler alert for later. But one of the ideas that I had is what I think is there isn't a venue yet for artists like AI artists. Let's just create its own category to really show their work in any formal way. So I have a project that I'm working on in the background, where I'm trying to pull together the AI art awards, and it would literally be that. So artists can, what they would have to do is they would have to obviously submit their image that they've created, but they'd have to submit the prompt and all the metadata that goes with it.


But the idea would be to take it from the screen into a physical piece of art so everybody could enter. We then have some very well respected judges from around the industry to come and judge the entries. We get down to say a short list, maybe, you know. Maybe then we put it to a public vote or something. I don't know. We haven't worked out the details yet, but the idea is that ultimately, the winners and the different categories, the art would be printed on canvas or like a photo and framed and hung in a gallery where people can come and actually see it as art right, and not just on a screen somewhere, and so it's not just locked away. It's a little bit more accessible for people and I think that would give it some legitimacy as well, because, like you said, there is a huge amount of creativity in it and They'll all be unique, because that same prompt will never give you the same image twice. It will always give you something different, and you probably will know this, and people who've been listening to my podcast for ages will know this, because I talk about it all the time.


But you know, the reason the systems are so good at the minute is because they have randomness forced into the system. So it's why you can ask the same question and you never get the same answer, even in chat, gpt or whatever, because there's a randomness factor that's in there, because if you don't have that, it doesn't feel human, and so I think that could be really interesting, because you know somebody gets something they're really proud of, they're really happy with. They've used probably a very complicated, you know some sort of prompts to get that Like. We could have like a, you know, you have one category around


you know black and white, I don't know profile pictures or something. Do you know what I mean? And it's like you would get photorealistic, you know, sort of imagery out of that, but you could also have, you know, something that looked more like paintings or whatever. So, anyway, I totally agree with you and I think that that is a skill, and so I am trying to pull something together for us to be able to recognize some of the artists who use that as, or AI as, their tool.

23:43 - Robert Plotkin (Guest)

I'm curious to know in the criteria for evaluating the art, are you thinking of evaluating the end product only or also the process that went into it, or would those be separate categories in the evaluation?

24:00 - David Brown (Host)

That would probably be. That's a great question, by the way, and I'm not sure I've thought about it that deeply yet, but I certainly think that that should be a category. You know, that would definitely be a category I would see, but I would think we'd want to judge that. We would want to judge the art like you would judge anything else. Like if we say, okay, we want you to generate photorealistic, you know street photography or you know whatever, then that would be judged like it would be judged in any other competition.

24:31 - Robert Plotkin (Guest)

Right, I mean it's. It almost strikes me like maybe it's analogous in the Academy Awards to you judge the pictures and the best picture and then you judge the script writing and then you judge other aspects of what went into the final product, not just the final product that's, and typically we don't do that with art. Right, you have paintings and you just evaluate them yeah. This provides an opportunity to delve into what was the creative process that went into it and evaluate and reward and recognize that separately.

25:04 - David Brown (Host)

So, yeah, anyway, spoiler alert, we will see that later this year where we're where I've found someone who I've been talking to about it, who owns a gallery and who's in that world and that also thinks it's a great idea. So she and I are working on that and hopefully we're going to reach out, see if we can get some sponsors and if we can get some sponsors and it'll be off and you can look for that later this year. Anyway, that was a yeah, I'm glad you mentioned it because it was a nice little segue for me. So just cheek at least and look at it there. Going back to something you said way closer to the beginning, and I'm not sure if you answered it, but you mentioned that you didn't think that we were asking the right questions. What are the questions we should be asking?

25:48 - Robert Plotkin (Guest)

e of a TV show from the early:


You put a lot of effort into creating each instance of a work because a lot is writing on it. You used to do that with cameras. You had 12 or 24 shots. You didn't just snap pictures willy nilly. But once digital photography came along, especially as storage space went down to basically zero cost, what do you do now? Snap, snap, snap, snap, snap, snap hundreds, thousands of photos and then you evaluate and you curate. So it moves from design first and then evaluate and revise to create first and then curate, evaluate, maybe revise and then go back. It flips the process on its head. So it actually requires quite a different set of skills, somewhat like the difference between a musician in an orchestra and the conductor. I hate to use sometimes terms like low level and high level because it has a evaluative connotation to it, but I would go from maybe artisanal or constructive to basic yeah, basic to fundamental, it's the fundamental skills right Trimitance.

29:06 - David Brown (Host)

It's the fundamental skills when you were talking about art. If you're a painter, you have to have the fundamental skills to understand how to mix the paint and how to use a brush and how to do the technique. You have to understand the technique of how to get the paint on the canvas, for example. Or if you're a sculptor, you have to understand the techniques to make the shape that you want. It's not even about understanding what shape you want to make, which is a whole different kettle of fish, but there's that fundamental layer. It's the same with musicians, right? You need to know all your chord shapes and you need to know that you have to have the physical dexterity. So it's the fundamentals you have to have, I guess.

29:44 - Robert Plotkin (Guest)

Right, that's right. So AI and computer technology, any technology that automates the process of creating works or parts of works, both enables and facilitates the use of these compositional, curative and iterative skills, like going from being a writer or an actor in a film to being the director, or even, nowadays, maybe, being the producer. So those are the kinds of questions that are really fascinating to me is, what are the skills that we need? I'll say for me. I know I'm a writer, much more so than an artist, or I'm also a bit of a musician, but much more of a writer. Until fairly recently, until around the launch of Chat GPT, most writers were still in the realm of those fundamental skills. To create a written work of any length, you had to write the words yeah Well yeah.


Whereas artists and musicians and other types of creatives had at their disposal these tools that let them take existing pieces and put them together in compositions and all of that. So Chat GPT lets you create those chunks a paragraph, an essay, something of that scale almost instantly, based on your input, and I think writers are now just starting to catch on to the implications of that that photographers and artists and musicians have been dealing with for decades. That, to me, is fascinating, and it's something I personally have been experimenting with.

31:28 - David Brown (Host)

Yeah, it's interesting that you say that. Do you think that's also why it's potentially or it's already having such an impact on business? Because so well, one half of business has operated using machine learning for decades, since the fifties. So you know, and, and frankly, a lot of the AI algorithms that we use today were developed in the fifties. They just didn't have the compute power to make them work the way that they knew they would work. But that's a whole different thing. But it's the difference in the types of tasks that it can do as opposed to what it used to be able to do as well. I think that's that's, and you're right it's.


It's the writing bit, because that's been such a fundamental part of business and tasks, and certainly things like copywriting and marketing and advertising and sales and all of that sort of stuff, where there's you have to have that. You have to have the written skills to be able to get to craft a message and to get it across, and there used to be a very limited number of people that had very good skills at that. Now everybody is above average at that. It's you're not, it's not as good as the best marketer and it's not, as you know, any AI tool is not as good as the best author and it's not as good as the best copywriter, but it is better than probably 90% of the general population at that task, and that's what's scary, because it it enables people who don't have that skill to now up their game to such a level that they could have never retained on their own, and I think that's the scary bit.

33:04 - Robert Plotkin (Guest)

Yeah, I think, from a business perspective and the job or career perspective, it forces you to ask okay, although, as you said, ai, like Chat GPT, isn't as good as the best writers. How many business tasks require writing that's at the level of skill of the best writers? A lot of, a lot of tasks don't. For quite a few years now, you know your weather and stock apps have been showing you textual descriptions of the weather forecast that look like they're written by a human, but they've been AI generated, because those are very narrowly tailored tasks that even before Chat GPT could be delegated to AI. It moved from just today's high and low to a paragraph you know that was customised for every zip code around the country, around the world that flowed. And the same thing for stock again stock market forecasts or summary of what happened in the stock market during the day. Those used to be things that you know. A news agency that was, let's say, in the US, had to have writers writing all that stuff every day and was then able to dispense with that.


So, from an individual point of view, it really pushes all of us who write for our career to ask what is it? Where is it that I provide value? How can I continue to be valuable and provide value in the backdrop, where the backdrop is AI tools that, as you said, can write as well as the average writer, I mean, the short answer is I need to be a lot better than the average writer and, honestly, I'm going to need to be able to learn how to maximise my leverage of AI tools to help me in writing. Both of those together will continue to make you valuable. And then the next thing is to just be a continuous learner, because whatever skills you learn today to leverage AI will rapidly not become obsolete, but you're going to need to keep updating them as the AI gets updated and improved.

35:14 - David Brown (Host)

They'll be different next week.

35:15 - Robert Plotkin (Guest)

Yes, absolutely.

35:17 - David Brown (Host)

I think, touching on something about the writing as well, I think there's a big part of it that's that has to do with topic, because there's a gentleman that sits behind me over there and he's an author, but what he writes about us, he writes about his experience of being multiracial back in the 70s and you know, in the UK and going to football matches and and what that was like. And he said you know, there's a lot of things that AI can write about and he's played around with it a lot, but he's like it can't pull out the real life experiences and it can't. You know, that's where it sort of starts to fall over. But if you needed to write something like an instruction manual or something. So I think what we're going to see is is there going to be certain types of tasks that it's really good at? But it's the emotive, it's that emotive, it's the human side, it's the, you know, it's, it's that sort of the emotional intelligence type stuff that it's it's not going to be able to do for a while. I think eventually it probably, you know, some AI tool will get to the point that it will be indistinguishable and then this gets into.


We were talking about empathy. I was talking about, you know, empathy from AI the other day and I said well, if it writes something, if you thought that was from a human, you would think that it was empathetic, and and if something had happened to you, then you would, because empathy is in the eye of the receiver, right? So if, if you perceive that as empathetic, then it's empathetic. It doesn't matter whether an AI wrote it or a human wrote it. And then they were pushing back and they're like no, but it can't be by definition, it can't be because it was done by AI.


And I'm like but it's like art, like art is in the eye of the beholder, right. And if I think something's art, I can look at a canvas in the tape with one. It's a totally white canvas with one red stripe on it, and they want like 50,000 pounds for it, and I'm like I totally literally could have made that myself. But but I didn't and do you know what I mean? And say it's, it's, it's all in the eye, and and I think so, but I do I do think that you, as a writer writing about certain topics, what what you bring is that human experience to it, and I think that's really where where the humans are going to make a difference and we're still going to be important Marketing copy or product descriptions or anything like that. Frankly, let the AI write it.

37:49 - Robert Plotkin (Guest)

Yeah, absolutely.


This is if you look at it as the AI provides a baseline of the average.


It's never I shouldn't say never, as you said, but at least now it's not going to be able to add the unique part, whether it be an insight or a part of your personality or your style of communicating or, in the case of your colleague, your personal background in your own life. It can't add that for a couple reasons. One, it just literally doesn't know about those things about you. And two, they're not part of its training data and it relies on its training data. And actually, the third thing is that at least the most commonly used forms of AI cannot use chat, gpt now they tend to provide you with what is the most statistically likely outputs based on the training data. So they tend to be, you know, think of it as providing a very great. The old term grey goo it's going to give you the average, which is not going to reflect that unique aspect of you. But that two things. There are lots of tools out there now that you can train on your own writing.


So, if you imagine, you know your own personal autobiography and it both not only had information about your life, but it was written in your writing style. And then you give it lots of other stuff you have written again to give it information about things that are relevant to you and about how you write. These systems now can do a decent job of then writing something new like you, to sound like you in your own voice. And then I understand a lot of successful professional writers have raised a concern: what's to stop someone else from doing that with my work? I mean, I was taking all of my Stephen King's work, you know, and putting it he's got tons of it right, putting it into a personalised AI and generating the next Stephen King novel from it. I mean, copyright law might come in to stop that, but technologically there's nothing stopping it from happening.

40:03 - David Brown (Host)

Absolutely right. But I think the again, what's different is that you can run a model, you could train a model on your laptop with all of Stephen King's material and use it to generate something and no one, like literally did. And if you deleted it afterwards, no one would ever be able to tell. And that's the difference. The use to that kind of thing used to be so expensive and it required so much power that no one could do it. But now we've got it on a laptop. I mean I have a relatively new, not a brand new, but I have a relatively new MacBook Pro. I mean it's insane the amount of processing power that I have on that. You know I can render hours worth of videos and minutes. It's ridiculous.


And you know, if I wanted to build a just a localised model on my machine using some pre trained you know a pre trained model that kind of came in with some pre settings in it and then just started loading it with, like you said, stephen King or someone else Sorry, we're picking on Stephen King, but whatever. But it's perfect because he has, he got tons and tons of material out there and he also has a very distinctive style in the way he writes. It would be very easy to copy that Stephen Fry did something similar at the Kagex Festival or the Kagex event in London last year and he showed a video or he showed a whole bunch of different videos of him talking and he's like none of that was me and he said all they did is they took all the Harry Potter books that I read, so the audio books. They put all the audio in from that and then use that to train his voice and literally it almost sounds better than he does.

41:46 - Robert Plotkin (Guest)

It's crazy. It is crazy, Same thing isn't it. It's crazy, and so I've been. What I've been arguing for is that two things I think humans will continue to be valuable, because each of us brings something unique, and if we recognize that and then keep developing our skills at expressing ourselves and by and using a tool to augment our ability to express ourselves, will still be valuable. I'm talking about, from an economic point of view, of course we're all valuable as unique people, independently of that.


But there's a separate question which does concern me. That's on what you call the supply side. I do not know what the demand side will be, but will the will, the mass of demand of consumers really want or or require that level of individual style or personality and what they consume, or will they be satisfied with the, the average that AI can generate, or the average plus a little bit of average human skill? If that's the case in music or film or or anything else, it might be that the really skilled human creatives won't have enough demand for their work to continue to really succeed economically. I don't know. That's a very open question that's somewhat outside of the control of creatives.

43:20 - David Brown (Host)

It's it's, it's yes, it's a really good question. It is and it's. I was having a conversation with someone a while back and we were talking about. We talked about we're only rich people going to have robots was the context of the of the conversation because they're so expensive, and it was like, well, only rich people are going to end up with AI and they're going to end up with robots and stuff. And their position was this actually, they think it's going to end up the opposite. What's going to end up happening is that the super rich people are going to end up with real pets and real, you know, real people helping them and they're going to end up buying real art and they're going to end up buying real books. The people at the lower end in the economy are going to end up with the robot dogs and the robot this and the AI read news stories and the and the AI written books, because they'll be cheap and they don't have that level of exclusivity to them anymore, and so, and I thought that was a really interesting way of thinking about it and I'm, you know, I kind of never thought about it that way before, but I was, I was converted. I was like, yeah, no, I totally agree with you.


I could see in, you know, 75 years, 100 years from now, that you know you're super rich. They're going to have everything done by humans, because that's where the that's where the artisans are going to be. You're still going to have humans that are going to do stuff. It's like today. You've still got milliners who make hats or you know cobblers who make handmade shoes. They're still out there, but they're so prohibitively expensive that most people can't buy them. And no, most people wouldn't even know where to go to buy one if they could. As opposed to everything's mass produced and you're getting, you know, fast fashion and the lowest quality. You know the crappiest clothes that you can get today because they're cheap. So it's interesting, and I do think that there is space for humans. But what's? We are going to rise to the very top, but that's only going to be 10% of everything that happens and 90% of the rest is just going to be, like you said, sort of average, sort of content.

45:28 - Robert Plotkin (Guest)

I mean, what you're describing, which I think is a definite possibility, is a return to patronage, which is, you know, think about Michelangelo, right?


I mean he benefited from patronage, which was the church, or the Medici's, or you know well the institutions or families that could pay him, at the top of his field, right to produce his, his art, and it's a strange confluence of technology and economic factors in the industrial age that have led to mass produced creative works reaching a mass audience, and there's nothing preordained dictating that that is going to continue. What you've described is like a return to patronage. It's something that journalists have been struggling with for a couple of decades now. As the demand for paid journalism has gone down. In the face of the internet making news available essentially for free, what is the economic model for a newspaper or journalism?

46:37 - David Brown (Host)

That's a whole other podcast.

46:39 - Robert Plotkin (Guest)

You know. But I'm saying you know people who have the biggest news organisations are continuing to struggle with this after 30 years of the internet. You know, I do wonder whether one of the end games of the New York Times versus open AI lawsuit is going to be a big deal, a large economic deal where the New York Times, you know, gets ongoing payment from open AI, which is a different business model for it than traditional, which has been advertising based. And I say more power to them if they can do that. They have to evolve with the times. I know that.


at the local sewage plant and:

48:07 - David Brown (Host)

For sure. I think what's interesting about the publishing industry is I think what really happened there is that they when they missed the boat is they missed the boat when the sort of Craigslist came out and gum tree. They should have bought those companies straight away, like they could have bought them in their infancy. You know they could have paid a ridiculous amount of money for them just to keep them under control, because that's where the majority of their money came from was from classified ads and all those smaller ads that you know they charged, you know, one or two dollars a week or whatever to run.


I've got a green sofa for sale, somebody can pick it up, and I used to work for the newspaper licensing agency here in the UK and I've had some chats with people in the business about and they thought that as well that because that completely eroded the foundation revenue on which those, those publications were built and that forced them to then go into this digital advertising model that has now just driven down everything.


It's driven down trust, it's driven down prices. You know nobody buys physical paper Like it caused a whole bunch of problems. But and you're absolutely right and you know, I found it really interesting, in particular the New York Times thing. So they filed the case against open AI that was pretty widely publicised and then they announced they were coming out with their own AI and I just thought, okay, this reeks of hypocrisy a little bit, in that they're. What they're trying to do now is they're just trying to protect their investment in their own tool that they're doing. So it took a little bit of the altruism off of it for me, I have to say, just because I was like right, so this is actually just more about business and money than it is about anything else, I think.

49:57 - Robert Plotkin (Guest)

Well, you know, if anything, I give them credit for then reacting more quickly than they did in the early days of the Internet, as you said, all the major news organisations. They just kept pushing back and saying we're going to keep doing things the old way, and even denying the fact that consumers would welcome the new way. You know they're like no one's going to want free content, it's going to be too low quality. Who's going to want that? People are going to still come to professional journalists and pay for it. And they were wrong.


I mean, it always amazes me, and I think this this was the innovator's dilemma right which is that you always have large incumbent companies who have built up empires on a certain way of doing things and then when a new, radical, disruptive technology or business model comes along, often they find it hard or impossible even to recognize the disruption of this new model, much less to adapt to it, and then that creates opportunities for new. I was just thinking recently, you know, about Amazon, because everywhere I go I see an Amazon truck going around. I buy tons of stuff for Amazon. And let me think about, you know, amazon conducted a takeover of the retail industry in plain sight of all the major retail players for 25 years, and yet I don't want to say that the Walmarts and others of the world didn't try, but they were so slow and so much in denial that it didn't matter that everyone knew what Amazon was doing.

51:30 - David Brown (Host)



And that and Amazon's still better than most of them. Frankly, I don't know. In the US I haven't. I haven't lived in the US for 25 years now. But in the UK there are a couple of retailers who really have stepped up and who do a very, very good job. They still have physical locations and in some shopping centres and stuff like that so you can go in, but they're generally very small and they're mainly just to handle returns and you know people go online. They've made it extremely easy. They carry loads of different brands. For example, they carry the gap. The gap doesn't sell in Europe anymore. What they did is they basically sold the rights to use the logo and the branding and I think they supply some, some of the clothing and stuff like that, but basically the gap just completely left. I think they're only in the US now. Don't quote me on that, but they certainly aren't here anymore.


But you can buy Gap, you can buy Nike, you can buy Adidas, you can buy, you know, name brand stuff through this department store that doesn't sell that stuff on the floor. So you know, they reacted, I would say, pretty quickly and have sort of sorted out their business model so that it works for them. But there's tons and tons of other ones that are like you said. You know, they just never, they never caught up and it was way too late by the time they tried to do anything.

52:51 - Robert Plotkin (Guest)

You know, and so it's a lesson. Let's take it back to the individual creative level, which is it's very easy and I relate to it. It's very easy to be in denial about the change that's underway as a result of generative AI. Let's say for writers, from large language models, from artists, from generative artwork tools and to say the things that we just kind of joked about the newspaper saying 30 years ago no one is ever going to want that low quality content. People are always going to come to people like me because I have a level of skill that other people don't have, and you know what? There might be some amount of truth in it, but the question is, how much a truth is there and there? How much will someone need your skill? How much will they pay for it? Relative to other things, well, as you said, there's still a demand for cobbler's, but quantitatively, how much demand is there for it?

53:54 - David Brown (Host)

Much, much lower so on an individual level.

53:57 - Robert Plotkin (Guest)

I think what I'd encourage people to do is to try to become aware of ways in which you might be in your own personal denial.


Open your eyes to what's out there. I suspect if you're listening to this podcast, you know you probably already are, but really start learning how to leverage the technology, how to make use of it to advance yourself now, although it might seem like we've already advanced a lot, we're still in the early days. There still is a lot of time to make use of it. But you know, if you are actually think there's something really beneficial about being someone who lived in the pre generative AI days, which is, if you have, let's call it, the old school skills and you can marry them with the skills of leveraging generative AI, I think you'll be more valuable than someone who, let's say, grows up now only knowing how to use generative AI, who doesn't know how to use a paintbrush literally or figuratively, because there is that need to revise, touch up all of the skills. If you can marry the two together, you will be super valuable, I think.

55:07 - David Brown (Host)

Yeah, I totally agree. So that brings me on to the fact that I'm conscious of time Now. We talked about in the sort of green room chat that we had ahead of time that you mentioned, that you know, and you've mentioned in the conversation that you've used AI in your own writings. So I'm curious. I know you have a book coming out soon, so did you use AI to help you in that and how do you use it to help you when you're doing your work through the authoring process?

55:34 - Robert Plotkin (Guest)

Yeah, so at a high level I and so the book I have coming out. It's called AI Armor. You can go check it out at AI Armor book dot com. I hope you don't mind the plug.

55:46 - David Brown (Host)

There'll be a link in the show notes, don't worry.

55:48 - Robert Plotkin (Guest)

Great. It is a book for founders and executives at growing AI companies about how to make use of intellectual property to protect your AI innovations, and so I drew on my experience as a patent attorney to write the book. It describes the strategies I've been using with clients for 25 years and that I've developed through a lot of experimentation and practical legwork on the ground with clients, so I brought a ton of personal experience to it. That's not part of the the ether, that was the personal part, and I did a ton of actual old school writing. Where did I use generative AI brainstorming? Ok, really good for brainstorming. I would know a general topic I wanted to cover in a chapter or a section and I'd ask give me a few different angles for this.


So let me just say one thing I'll say is I think a lot of people use something like Chat GPT to do the writing. They jump straight down, ask it right An essay, write a blog post, write an email. I like asking it questions where the answers help me to ideate. To stay at that higher level, first, yeah, what are some different titles I could use for this section? And then, if it gives me a title that resonates, I might go through. Third, I might say, no, I don't like any of those 10 titles you gave me. Give me 10 more, and then one of them will resonate with me and I might either run with it or revise it or say you know, I like that one, here's what I like about it, here's what I don't like about it. Now, based on that, can you come up with some more? So this is a kind of collaborative process. And then that title might reframe how I'm thinking about the section, which I may then write manually. Or I may start writing it manually. And there's a really valuable thing there's a lot when I get stuck. I'm writing a paragraph manually and then I just get stuck and I don't know what the next sentence is. So I'll feed in what I've written so far and I'll say what are some possible ways to continue. I love that. Everyone as a creative person knows that when you break the flow it's really painful and frustrating, but if you can keep those juices flowing it's just so much more enjoyable and productive and efficient.


Sometimes I'm writing and I'll describe a concept and I want to give a concrete example of it, either theoretical or from history. I used some Chat GPT a lot for that. I would say small companies often benefit in competition against big companies by having patents that they can use. I know some of those examples historically, but also give me some examples of this. I wrote an article called Why I Use Chat GPT to Tell Me Things I Already Know. That's a good example. I mean, I know that small companies have used patents against big companies. I've had my own clients who've done it. There's some big historical examples, but that doesn't mean that those examples are always at the tip of my tongue or at the top of my mind when I need them when I'm writing. So I might ask Chat GPT, give me the examples and I'll say to myself oh yeah, I knew that.


I just didn't remember it at the moment and again, help me keep the juices flowing. I'll stop there for now.

59:28 - David Brown (Host)

I was just going to say it's similar for me. A lot of times I'll say I'm writing about this topic, can you give me an outline, or whatever, and a lot of times it will come back with stuff that I didn't think of and I was like, oh yeah, that's a good point.

59:44 - Robert Plotkin (Guest)

It's like I hate that this thing's smarter than me half the time, but never mind, it's not any different from having a writing assistant, an intern, a collaborative, maybe even a peer who you bounce ideas off of and most people don't have any shame in that, but yet it's interesting it does in most people, I'll agree the same with me Elicit a little bit of a shameful reaction. Sometimes it hits our ego, maybe when it's an AI that does it that wouldn't if a human did it. So I have it. Come up Examples high level ideas and changes in tone. I'll say it's not that great at humour, but sometimes I'll say I'll write something and I'll read my own writing and say that's too stodgy.




I'll give it a whole paragraph or section and say can you come up with some ideas for making this more dynamic or personable or some other style? And again, one high level point I want to make is sometimes it'll come back with something and I will throw out 90% of what it comes back with, or I won't even use any of what it comes back with, but reading what it wrote just jogs my own creative juices to help keep me going. So these are all ways in which I just want to stress to people that using the tool to do the work for you is not only often not the best use of it. There's so many other benefits that don't involve using it to do the work for you that help jog your juices, jog your memory and help you move forward to be a better writer yourself. 100%.


Somebody suggested we should call it augmented intelligence, not artificial intelligence, which I think is a good way to express that, and I think it's also funny that loads of people would be quite happy to use a ghostwriter, but then they want to complain if someone uses AI, which is basically the same thing, yeah, and honestly, if you're interacting with AI in this very fine grained way, you're probably getting something out that's more you than you would with a ghost writer, or often what do you give them a high level outline and tell them this is the topic, and then they kind of churn out the whole thing for you and maybe you revise it?


Yeah, Robert, thank you very much for your time. Today. I'm going to work. We're over just over an hour now, so I will let you crack on with your day. Tell everybody I mean we've mentioned the names of the books and all the everything will be linked to. Everything will be in the show notes. I've already put the links to the Amazon books and everything in there. I don't they're not affiliate links, so anybody can just click on them and just go to them. I don't get anything for it and I'll put a new link to the website for the new book that's coming out. Is there anything else where people can go to find you or where they can go to maybe get you to help them if they've got IP issues?


Yeah, thanks so much. The other two places I post really frequently on LinkedIn about developments in AI, specifically though generative AI, its use in writing and its relationship to patents and intellectual property. So you can find me there LinkedIncom, slash in, slash Robert Plotkin, and then if you are at a company that's developing innovative AI or software technology and you're interested in intellectual property protection for it, you can go to my law firm website at blue shift IPcom and contact us there or schedule a free initial consultation.


Perfect, robert. Thank you very much, it's been amazing. It's been a great conversation.


I really enjoyed it. David, Thanks for having me.


Thanks, I'll speak to you soon, bye, bye.




More from YouTube