Inside The EPO’s AI Push With Angel Aledo Lopez

What happens when the European Patent Office’s CTO and COO says the hardest part of technology isn’t the code, it’s the culture? Lee and Gwilym sit down with Angel Aledo Lopez to unpack how AI is changing the way patents are examined, filed, and managed, and why human judgment still anchors the system.

Transcript

Wordplay And Weather Banter

SPEAKER_02 0:00

Hey Grillum, how are you this? I was gonna say this fine day, but it's it's one of those days that's shifting between sort of overcast and sunny at the moment. You don't know quite what to wear. But I'll see you're you're in shirt sleeves, so you must be somewhere nice.

SPEAKER_01 0:12

I mean well I'm indoors, but I've just I've just cycled and been blown backwards all the way from central London, which is just depressing. Not literally. Is it windy out there? Yeah, it's awesome.

SPEAKER_02 0:22

Yeah. No, uh I've not ventured out much. There we go. How how are you?

SPEAKER_01 0:27

You good? Fine. In fact, it brings me to the topic we decided we'd discuss here. The wind was so loud that in my ears there was a cacophony. A cacophony?

SPEAKER_02 0:37

That's a lovely word for much that's much underused, I think. You were teasing me for saying bailiwick just now, but I think these are bailiwick. I I love I love bailiwick. Bailiwick is one of those words that should be used more often.

SPEAKER_01 0:51

So I've got I've got four written down here. I've got bailiwick. Don't know what actually what does it mean?

SPEAKER_02 0:56

Is it to do with castles? So I mean, I know the um Guernsey is a bailiwick, isn't it? I know I know that much that the jurisdiction of Guernsey is a bailiwick. But um I won't I won't profess. So other than kind of uh a contained perhaps legislative area or something like that, don't know. You would think you would think so, yeah, you think about modern Bailey castles, aren't you? So maybe it maybe it does derive from that. Yeah, yeah.

SPEAKER_01 1:20

Maybe maybe then there's obviously cacophony, it's a biggie. Yeah, uh, which I think was under less probably less underused, but great. Um, especially if you don't know how to pronounce it. I I think it's cacophony. It could be cocophonie, I don't know, but cocophonie, yeah, yeah.

SPEAKER_02 1:32

No, I think you're right, it's cacophony. So come on, you've got two more, what are they? Because uh our guest is champing at the bit to get on the podcast, and you're smart proceedings at the moment. We're meant to be having banter.

SPEAKER_01 1:41

This is the banter. Come on. Banton, bant on. I'm banting. Hoi poloy. Hoi poloy. Hoi poloy. That for a long time. Yeah, I think we should use that more. And I've decided my favourite one I've never used is embroilio. Imbroilio? Imbroglio?

SPEAKER_02 1:59

Oh, I don't yeah, not not a word I'm familiar with.

SPEAKER_01 2:02

I think I think it means if you get it kind of all mixed up in a really complicated set of affairs, it's an embroiderio.

SPEAKER_02 2:08

Oh, okay. So you get I've I've heard that you get embroiled, but maybe that's from the same route. It's got a G in it.

unknown 2:15

Yeah.

SPEAKER_02 2:18

Well, I enjoyed that band for me. Yeah, yeah. I I used a word in the office the other day and no one knew what I meant. And for me, it's just such a normal word. So I'd I'll test on you. It's not a it's not kind of a highfaluting word that you've that's not used often enough, is it? It's not a highfaluting word like like you've just used. It's the word squinny.

SPEAKER_01 2:36

Have you ever come across the word squinny? Yeah, it's the little washer that goes under the nut on a kid's um swing. Is it really? No, I've no idea what a squinny is.

SPEAKER_02 2:47

Have you got no idea? So in my world, it's someone that's excessively complaining or moaning. And uh so back in Portsmouth, we would use it to say that someone is like behaving like a baby, you're being a squinny. And I used it in the office, and about 15 people looked at me as if to say, you've just made a word up.

SPEAKER_01 3:02

Yeah, there's a word like that up um Nesh. Have you come across Nesh? No. Uh Nesh is a northern word that I picked up from Manchester Current, meaning that you feel the cold lots, and you're gonna have to wear too many coats and things. That'd be Nesh. Oh, we could do a whole podcast on this, couldn't we?

Guest Introduction And Background

SPEAKER_02 3:18

I'm writing it all down. Yeah, we that we could have a veritable liturgy of words. All right, all right, all right. It seems terribly unfair, doesn't it, when we're going to introduce a guest for whom English is not their first language. It seems terribly unfair that we've started the podcast on a series of random words that they've probably never come across before. Angil, welcome to the podcast, sir. Very good to have you on. How are you?

SPEAKER_04 3:43

Nice to meet you, Lee. I think, yes, for me it has been difficult to follow up. Yes, sir. Um, I mean, the weather as well is very good here in Germany. I am in Munich right now, sitting here, very sunny today, so I cannot complain that sense. But yes, I mean for me has been difficult to follow up that conversation.

SPEAKER_02 4:01

We're trying to use plain English from this point forward.

SPEAKER_04 4:04

Please, I would appreciate that.

SPEAKER_00 4:08

Lee Davis and Willem Roberts are the two IPs in a pod, and you are listening to a podcast on intellectual property brought to you by the Chartered Institute of Patent Attorneys.

SPEAKER_02 4:25

It's always nice for our listeners to get a feel for who our guest is. So would you care to spend just a few minutes talking about you, who you are, what you do?

SPEAKER_04 4:34

Yes, yeah, I will. Well, my name is Angela Redo Lopez. As um maybe you can imagine from my name, I am from Spain. And I am right now the chief technology officer and at the same time the chief operating officer at EPO, the European Patent Office. So it is a bit strange that we are combining these two roles, but I can tell you, and we maybe we can discuss about that later. It is also a huge opportunity for us to look into the two aspects of the implementation of tools and AI, the operational aspects from the support to examines and formality offices, and the IT challenges that the tools can bring as well. So quite a challenge, but a very nice opportunity. I have been here at EPO already seven years. Um and before that, I have spent 17 years in in Alicante, in the European Union Intellectual Property Office. So basically 24 years of experience in the area of IP in general. Uh, seven in patents, uh, 17 in trademarks and designs. Always supporting the implementation of projects and the digitalization uh uh of the office. This is my background, and of course, I have a technical background in the area of IT. But as I said, nowadays I am more focused on operations than anything else.

SPEAKER_02 5:57

So without being very rude and asking age questions, how long ago did you get into IT?

SPEAKER_04 6:02

Uh I don't remember anymore. I don't know. I I have been always working in that area. Even when I was a kid already, I was playing with computers. I remember my brothers that they were playing games with uh a very old Astran CPC 6128 or something like that model. And instead of playing games, I I was doing programming myself. I was reading basic and everything. So I I mean that has been my passion always to to work with computers.

IT’s Real Constant: Change Management

SPEAKER_02 6:31

Because I've always I've always thought that people that have worked for uh I have a couple of friends who've worked in IT probably since they left school, so that would make it 40 years in in my world. And I always think it's quite interesting to have conversations with them about how their world has changed, because that's it's probably the part of the world that's seen the most changed, uh, probably at the greatest pace of any of the areas that we work in.

SPEAKER_04 6:54

I okay, I have a different interpretation of the facts. If you want to call it like that, yes, this has been changing all the time, but uh uh there were changes as I am telling you. When I started programming, there were a couple of programming languages, then older programming languages came. Some of them are already deprecated, the people don't use them anymore. But what I like from I from IT and from the IT part is that um what I have seen is that uh in independently of the use of different technologies, programming languages over time, uh what I see is that IT or technologies having an impact on the way people are working and interacting between themselves. Okay, and this is from my point of view, what has been consistent over time. I have seen the first times of uh or the initial times of uh digitalization of organizations or companies, universities, or somebody started in the university already, a number of projects there. And what I have always seen is that independently of the of the technology that was available at that moment in time, the challenge was always more the change management and the and and the other aspects that were not the technology. Technology has been always there. Doesn't matter if you are programming in Java, Python, BASIC, something else, the key challenge is always how to deliver value to the business, how to transfer an organization and the business model on the basis of what technology can do. And I guess that this is important as well in the in the area of AI. So, yes, some people can argue that technology has been changing a lot. And it is true that, for example, AI has been always there, but the way we have been approaching AI or implementing AI has been evolving over time. But still, the challenges have been always business challenges, I would say.

SPEAKER_01 8:39

Where do you want to start? I Lee wanted to start at the beginning. We've heard the kind of the early days, everyone else was gaming, and Angel was decompiling things and changing and writing code and stuff. But um, I I gathered from actually Steve Rowan, by the way, um recommended we talk to Angel and he said you've got to ask him about chess. So let's hear a little bit about the chess factor. And you mentioned that that was part of what got you into AI, I know.

Chess, Deep Blue, And AI Curiosity

SPEAKER_04 9:02

It is, it is part of the history as well. As I am telling you, I had that uh background. I I I was working a bit on IT, I like programming, I was exploring a bit uh technology without uh still being in the university, but I was also at the same time a chess player. And at that moment in time, when I was still learning about everything, I remembered that for us, for me, as a chess player at that time, Kasparov was the best player in the world. Okay, and I mean I never imagined that uh uh a machine technology, coming back to was not only the machine, but software and solutions there, that they they were they were able to beat someone like Kasparov at that moment in time. So I remember very well the times of uh Deep Blue. I don't know you you heard about that machine. Deep Blue was the first machine that was implemented using a kind of an AI system, but also sank hardware behind, so it was not only an AI solution. And just at that moment in time, the people, chess players like me, we never thought that that was going to be possible. I can tell you. I was, I mean, completely we we we took for granted that uh he was going to win that that match, and he lost. Uh, he was that was a surprise for him as well, I can tell you. And and there are books and uh and movies about that. Uh and for me was uh something that's a wow, I mean, if uh if Kasparot, that is was kind of uh, I mean, for me, that the best player that we have ever had in history, they are in chess. That is a machine, that is a tool, that is a technology. I didn't know about artificial intelligence yet, or at least a part of the books, nothing else more related to implementation. If technology is uh good enough in order to beat uh Kasparot in chef, that I I thought that it was the most complex problem in the world as well at that moment in time. What can we do with that in other areas? And then it's when I started my passion and then my let's call it my training as well in that area of AI, because yes, I wanted to learn, I wanted to understand how this was working because I couldn't figure out myself how that could be possible.

Centaur Professionals And Human Oversight

SPEAKER_02 11:06

So Angel, can I come in there a wee bit if that's okay? So um, as as Gillam knows, I'm quite passionate. I I am not a chess player, I've never been a chess player. I did it very, very badly at school, so I gave up. But in my adult life, I've become quite interested in AI and the way it impacts on the world of work. So I've I've I researched quite extensively Kasparov's experience, the whole deep blue thing. And so correct me if I'm wrong, because obviously I've I've looked at this as not a chess player, but I know that he was quite he was quite defeated personally by the loss to Deep Blue, and it took him some time to recover from that. But one of the things that I think he did in in recovery was to describe human players collaborating with computers in the world of chess. And I know that a hybrid game came along, so you could be machine against human or machine against human playing alongside a machine and or human against human. And I think the kind of knowledge was as the game would develop that the machine would win more often than not, when in fact it was the hybrid player, so the human playing alongside AI that became the more often victorious player. And I think he described that player as a cyborg or a centaur. I think centaur was the word that he used. So I quite often use that chess analogy when I'm I do quite a lot of public speaking on the membership circuit. And when I'm talking about the future world of work and what the future professional might look like, I always talk about the future professional being a centaur. So combining the best of the human and the best of AI together in in kind of like one entity, it's not it's not an AI or human, it can only ever be AI and human, and that that will be the most successful professional. Is that is that reasonable?

SPEAKER_04 12:53

It is, if we are discussing again about the patent system and the examin and the examiners, but if you look back into that message uh at that moment in time, that was the reality at that moment in time. Nowadays, uh no human can beat a machine. Okay. Nowadays, nowadays only machines can play with uh machines, chess, because I mean, even in a very in a device like your mobile or uh a device with very low computing uh uh capacity and memory, uh computers and machines are far better than us. So, no, we cannot compete right now. Chess. I am talking about chess at chess level. We cannot compete with the machines. At that moment in time, that was the case. And nowadays this is not the case. But I agree with you. I I like that analogy. I use that internally as well to explain to people that there are two possibilities of using AI, and this is very risky, obviously, the introduction of generative AI in our business and so on. Some people see that the answers from those tools can be very powerful. They look that they know everything, but there is a risk that people have brought all the responsibility to the machine and they don't integrate that uh that let's call it that uh uh potential or the power of the AI in the way they are working. I like that analogy in the sense that, of course, I see more examiners, formality offices, and people here in our office uh using the technology to support them and not the other way around, not using the technology to replace part of the things they are doing. Obviously, we can always argue about things in the sense that there are areas where the machine will be always better than a human. A human cannot speak 100 languages at the same time. Okay, we are not going to pretend that uh a human, a patent examiner, can translate any document, any prior, a prior document coming to the office in 100 languages. So there are things where the machine, in this case AI, can do machine translation far better than any human, and we need to use it as well. But still is the examiner the one that has to make a final assessment of all the documents. So we can use the AI, but the responsibility of the examiner is still there.

AI Use At The EPO And Policy

SPEAKER_01 15:10

When um I mean there's gonna be increasing uptake of AI, obviously goal of the EPO, not make a sense. But um, are we gonna see a time when it's EPO is gonna have to declare how much assistance AI gave, for sample in an office actions up something on the cards, do you think?

SPEAKER_04 15:26

Most probably, I mean, we are very open to share what we do with AI, obviously. And we are sharing even tools with the with the users, like the legal interactive platform. If you go to my EPO, you can find a tool that has been grounded with our legal text. It's using generative AI to give you answers or support in the area of uh legal uh support. We have these kind of tools internally as well. We use tools in many different areas from pre search to classification, machine translation, as I said, to support to examiners in in examination and so on. The level of use of those tools, as we have been discussing, and how much is the translation of the document relevant for that specific search? How much is the contribution of that specific translation to the overall outcome? I always tell people here the final outcome, and this is our policy, our AI policy says very clearly the responsibility of the final actions of this office will remain with our examiners. How much you can benefit as an individual examiner from the use of those tools will depend on specific cases, on specific technical fields. I can tell you AI is there, can deliver value, but we see that the level of quality in the different technical fields is also different. It's not the same chemistry than AI, than fields that are more based on images, for example, on drawings. So it is very difficult to say what is the level of contribution, that the AI has been integrated and is going to be integrated everywhere. I think this is right now quite clear. Okay. In our normal daily life, when I joined this meeting, I already got a message from Zoom telling me that AI companion is there to support me. We have it in Zoom, we have it in Teams, we have it everywhere. When you go to search to Google, you get already some kind of uh summary of the results coming from Gemini. So AI is going to be integrated everywhere. How much you are benefiting on a daily basis on a particular case from that use of AI, well, that depends on the case. Uh as I said for the time being, the machine is not there to replace anyone, for sure, that I can tell you. And we will see in the next five years or 10 years what is the situation. But I still believe that there is a role to play by a human in that in the assessment of the inventive step, in the assessment of uh if two claims are similar or not in this specific context. So that level of expertise is something that's good in our office, is that we have uh 4,000 examiners that are specialized on specific technical fields. For the time being, I insist machines are good, but they are good in general terms, not when you go to concrete specific technical fields.

SPEAKER_01 18:05

And that's interesting because I think um on my side as a European patent firmly, we in my firm we have clients asking us to trial certain office action response tools. And we obviously what we're currently doing is kind of quality controlling. So we'll generate something based on the tool and then we'll do a full human job on it and compare them. What we noticed is how quickly it's improving to the point where we begin to think, yeah, I mean, we could perhaps accept some of this with slightly less supervision. Um we're not gonna do that yet because we've got to talk to lots of insurance people first. But it's coming to an interesting time, I think, where you can. I mean, the fun kind of sci-fi outcomes, of course, AI generates a response, and an AI examiner comes back and says different things. But then of course, this is the same brain. I'm not too sure. I think you just work it out for itself, which brings lots of interesting implications. But I think the whole professional and probably the whole examining communities think what hoping that's not coming from why.

Can AI Invent And Patent Volume Shocks

SPEAKER_04 19:02

Yeah, well, uh, I think that uh that uh was already the argument sometime ago when we we were moving from a traditional let's call it system where we were working with paper and then computers came, and then a new search system more powerful with more data. So, as I said, the the level of digitalization, the impact digitalization has had already in the different offices has been huge. Yes, I can tell you that. Okay, I mean this office is not working like uh maybe five or ten years ago, distributing paper around between offices. Okay, we are in a digital world, we have access to new tools, and and that implies as well some benefits, and we have seen that over time, and it is not only because of AI, okay? AI is one to is going to be one of the players, of course. AI alone cannot do anything, you still need to deliver tools. I mean, if you go to Just GPT, you put two documents and you ask for an analysis, you can get one page of analysis, but still you are not going to be able to process that without proper tools to compare planes, to highlight things. So that is still part of a traditional IT development behind. And I can tell you that uh the last language models, generative AI, it looks very powerful, but it's not the solution to all problems. Okay, there's still the traditional and all AI is good for many specific use cases. As I said, um, I mean, we started uh that uh the implementation of generative AI tools here uh quite early in the process. So ChatGPT came, we saw the potential of ChatGPT. The the most fascinating thing of ChatGPT was that um, and I am mentioning them because they were the first one coming to the market, was that in the past we were training models for a specific task. I needed to train a model for machine translation, I needed to train a model for classification, I needed to train a model for search, I needed to train a model for something else. So every task, even if I was using AI, needed some kind of uh data training, process, tools, and everything else. That process was quite expensive, and also sometimes the results were not there. Suddenly, a tool came to the market. Well, I was able to translate to 100 languages, they were able to discuss about legal text, they were able to discuss about classification at the same time. So it looks like one tool can do everything we have been doing for years here with the training of our data, and that tool is accessible to everyone in the market without spending one euro almost at that moment in time. But it was not true. It was not true. I can tell you the quality of the classification is not the one we we can get from the tools that where we have created that content. Maybe they are getting the classification from every single office around the world. They are getting a lot of noise as well with this huge amount of data they get. So when you are asking, can you give me a symbol or can you give me the classification of that piece of text that claims or description? Chat TVT, I can tell you, or OpenAI or any kind of generative AI models can give you a Good answer sometimes. If you say this is not the answer I was expecting, they will give you another answer, don't worry. When you go to the traditional OEI that we still use in that field, you get a consistent answer from the tool because it has been trained for that specific task. So, no, I don't believe that these kind of tools are going to replace everything we do, and they are going to have that general knowledge and to be able to perform any task to the level of quality we need.

SPEAKER_01 22:26

Thank you. And actually, that possibly students is a cheeky question I'm going to throw in, which is a legal question. I know that that's not your day, not your Bailey Wick. That's a good English word, not your responsibility. But out of interest. Obviously, there's been a huge legal debate on the um practice side about whether AI can invent. And the I mean most legal authorities say it doesn't really matter because the law doesn't embrace that. But do you think AI is capable of inventing right now?

SPEAKER_04 22:52

Okay. I I think um we have seen already cases in the UK, but also in Japan, and you can go to the newsletters and find that, where there are very big companies that are using generative AI to generate patent applications. So how much contribution is there from a human to that invention or not, I cannot tell you right now. But the fact is that yes, they can, they can have a significant contribution to the invention. If you look in in, I mean, when people are talking about uh if the tools uh can deliver quality or not, or the level of maturity they can invent or not. Look into the Nobel Prize the this year, the previous year, okay? Uh I mean, there were two Nobel Prizes that were won by people working in the area of AI, and one of them in the area of chemistry, because they have delivered a tool that is able to predict uh things that were not possible by humans in a way that no human can do and will uh will be able to do ever in our life. And and this tool can deliver a lot of value to people, then to generate inventions, of course, and to accelerate innovation in that area. So, again, the question is more back to you, depending on how you define if the level of creativity and how much people are benefiting from this, for example, that specific implementation from DeepMind. I think yes, this tool is contributing to the invention. How much? What is the level of uh human intervention, invention there? And this is the part that is difficult to assess right now.

SPEAKER_01 24:28

And I mean I'm I'm gonna neatly dodge your uh deflection of the question, but I will I will comment interestingly. But um, I saw in Japan last year um a big one-known company, I think filed 10,000 patent applications in two days. Yes, um, as a demonstration of AI other than everything else is I've never seen anything like it, it's quite astonishing. And it raises so many issues about the noise that can be created, about the impact on any office of having that that surge on third parties we then have to work out where they stand in relation to all the bus. So it's gonna be interesting to see if people start not exactly abusing, it's not quite the right word, but certainly creatively using AI does make such a such a noise. Yeah, it can't work out as we stand. So that's not definitely not your problem. But it's an interesting, interesting one.

SPEAKER_04 25:21

I I can't tell you what is my opinion on that topic because yes, we get that question from time to time as well about the volume of work that could be generated by uh by artificial intelligence if we start having this kind of uh applications. Okay, for the time being we we haven't seen that team, but at the very end, you don't develop something, you don't create something, you don't innovate just to have a patent application. You need to have a business model behind, you need to have something to deliver. Okay. So, I mean, 10,000 applications, that's fine. But uh, what are you going to do with those applications? Uh, what is your business model? I mean, managing and maintaining that portfolio in the long term can be very expensive as well. It is not very difficult to grow very quickly in that sense, but then maintaining in the long term, we we see many, many companies that are trying to reduce their portfolio of uh patent applications to keep the the good ones, the ones that deliver value to the company, not to keep any potential application they may have. Okay. So I think um that will be also something that uh will influence a bit uh that creativity from people to use AI to generate whatever uh they can.

Interoperability Across Patent Offices

SPEAKER_02 26:35

It might be provocative. Might be provocative, might not be provocative, okay? Depending on how Anghell answers it or doesn't answer it. Um so and I know that the EPO is going through its own extensive digital transformation project. The UK Intellectual Property Office is doing the same. In fact, most IPO offices around the world are doing that. And so the the one in the UK is based on the.gov.uk platform has to be for kind of largely political reasons, which means it diverges from the system that the EPO is using. I guess my question is, in a in a world where interoperability is becoming more and more important, does this divergence of approach from patent offices in increase tensions in the kind of that the digital IP world? Or have we reached a stage where interoperability is straightforward and we don't need to concern ourselves about different offices having different systems?

SPEAKER_04 27:32

Well, I can answer that question for sure. I am hoping to have very soon tomorrow and the day after we have meetings here with our member states. So the UK office will be here as well. So we meet the 40 offices, I mean EPO and the 39 member states on site to discuss IT cooperation, so IT projects. Didn't know that. What what what timing? Yes, I mean that just in time. So we are going to discuss about AI as well and how the different offices are approaching the implementation of AI, their experience, and also of course about the technology or technological aspects. That information, I mean, we work together, okay? We know what the different offices are doing, they know what we are doing as well. I think we need to understand that you said that there are big offices like UK, Germany, France. They are not only working in the area of patents, okay, they are delivering services in UK to uh in a portal to that are covering from patent applications to more maybe you are um uh driving license registration and so on. So they need to also in the benefit of interoperability and being user-friendly, payment systems, identity management, and so they need to converge more to the UK system than to the patent system. Okay, yeah. So we need to understand those differences. They do that in the benefit of the users, okay? And of course, in the background, we work with offices, with IP5, with WIPO to uh to try to be as compatible as possible, like for example, using standards, resting the 96, XT36, standards to exchange data, standards to file applications, standards to, for example, the ST26, or for the sequence listing that has been adopted by all offices. So you know it you want to file sequence listing to UK, you guys go use the same format you will use in this office, or if you go to use DTO. So the uh convergence harmonization is in a way trying to we try to do that in the background, okay? But it is not always possible from the user point of view, and there are very good reasons to do that. As I said, many times the being compatible with other services that uh your country, that uh organization is offering is more important than than being compatible or being aligned with the button offices are offering.

Going Paperless And MyEPO

SPEAKER_01 29:58

I'll do actually point to the interoperability is the one and Leo actually kind of opened up the next question, but which is all about kind of the EPO digital transformation. I think Ansel, in particular, big plans go completely paperless. And I wondered kind of how that's going, what the response has been, what the plans are.

SPEAKER_04 30:15

In general, about the digitalization, obviously nowadays we have seen, of course, with that COVID situation and everything that uh happened, we we needed to move away from paper very quickly. We couldn't tell people you need to come to the office to sign that paper, otherwise, the the grant is not uh the decision to grant is not going to be sent. So that was done in a short period of time, I would say. We are now, you most right you know about that product, my EPO. We are now uh also the commission very old we had in the past, from the old mailbox to online filing, ELF, variable systems that were there for more than 20 or 30 years already in place have been replaced by a modern platform where we can offer more services. Now we have my IPO, we are integrating AI there as well with the legal interactive platform. We have a better way and the possibility to extend and create more business value, if you want to call it, and deliver better services to the users, to the end users. And internally, we are so we are going through the same transformation. We are getting we are uh step by step uh reducing the technical complexity of our systems, uh, creating also more modern applications and at the same time looking into how to integrate AI. The main thing that has happened, I would say, the last maybe 18 months, no more than that, is that the pace of development is going to faster than ever, okay? Coming back to that digital transformation. In the past, you could have a project that could last 18 months, maybe two years. You needed to gather the requirements, specifications, uh, implementation details, architecture of the system, and so on. Nowadays, with the use of these AI tools, the models are changing almost every week, if not every month. The APIs as well, the libraries behind, the frameworks. So the pace of development we have now, for example, uh most probably you are aware, we we have uh implemented, we are we we had a pilot phase for the oral proceedings supported by AI in both examination and opposition. So we have been delivering oral proceedings with the support of AI for the for the minutes of the meetings. The pilot was finished at the end of uh last year, and now we are moving to the towards the full implementation of that pilot. So this year is going to you are going to see that all oral proceedings OPO exam will happen using the support internally of AI tools. Well, we did that implementation, it was working very well. You don't have full control of what is happening in the area of AI. This tool is using copilot, Microsoft copilot in the background agents. Suddenly nothing was working when we needed to have an oral proceedings because Microsoft can decide at any time to change their copilot implementation, and then you need to adapt in a question of days, if not minutes. Okay. The same is happening with all these tools that are coming. We need to have the flexibility to adapt. We cannot just uh pretend that we are going to have an implementation today, and that implementation is going to last the next two years. We need to have a mindset where we need to understand that we need to deliver value as quickly as possible, but as well, the level of support and the level of effort that will be needed to maintain that application in the long term and to adapt to the changes that uh can come in that technology in the future is higher than in the past. So this is what we are facing now. We need to do projects uh shorter and in a shorter period of time than in the past, and and the level of effort to maintain and the skills that are needed are higher than in the past.

Fast-Moving AI, Pilots, And Ops

SPEAKER_01 33:50

And I mean, I think you you you're really worrying about acquaintances. So notoriously slow and also notoriously difficult to get right on launch. But actually, I just wanted to pass back probably on behalf of fashion. Congratulations about the transition that you managed to handle over lockdown, where you know, obviously those the whole world changed, and the shift to video conference film was rad with all of the discussion was about whether it was right or wrong to like nobody actually talks about the fact it just worked. I'm guessing that quite a lot happened behind the scenes while that was going on.

SPEAKER_04 34:30

Good guess, yes, of course. You can imagine the level of effort. Not only that, uh I mean I'm most proud you know about the EQE as well. Remember EQE when people had to come to the office to do everything in paper, and we had to we needed to manage, I don't know how many sites. Uh and nowadays it is uh everything done online. So, yes, of course. I mean, uh you can imagine the level of effort here to figure out, to challenge the system, to challenge the processes, to challenge to challenge the the way people have been working for a long period of time. It is not easy. I mean, again, I insist, it is not about the technology anymore. Technology is there, most probably has been there always. It is about change management, it is about challenging the way you have been working until now. And yes, I mean, I am personally very proud, obviously, because I was there as a chief technology officer. We are very proud as an office that uh that we have been able to take that challenge and to take some risk. Okay. I always tell people, of course, we can be always the last one adopting a technology, but that means that also the level of benefit most probably will be lower and that if we are early adopters. But that implies obviously a level of risk. So, how much risk do we want to take? What are the actions we can take to mitigate that? Well, as you have seen with the pilot of oral proceedings, let's run a pilot, let's gather some feedback, let's see if this is working or not, let's fine-tune the way we do the things with our users as well. And when we believe and we feel that we are ready to go ahead with a full implementation, then we do it. We did it with all the systems we have delivered until now. And I think this has been working very well. So, because that helps us as well to understand the needs of the use of the users and to see the limitations of technology, because as I said, technology is there, yes, but uh still with some limitations. So there are things that we would like to do and that they are not possible yet. So we need to wait until technology is there as well.

SPEAKER_01 36:27

No, it was um yes, I think you uh you're right that technology was there, right? I I get your point about change management. But the fact that you just sort of seamless actually helped manage the change. So I don't normally do agilatory statements, another big word there. Um I remember the debate very strongly, and there's a whole bunch of stuff about access to justice and being able to really do water proceedings. But a lot of it was what if the IT doesn't work? And if the IT hadn't worked, I think there would have been a lot more rollback and a lot more pressure. And what we saw instead was people doing it who really opposed it and said, Oh, actually, it was quite good, and I could see their faces and all that kind of stuff. So um no, I repeat that. I mean, I I think I would say a professional has a lot of confidence in what DPO does on the IT front because it seems robust. Um, I transitioned to the new filing process a year or so, and I'm I hate charge. And I didn't actually notice, I seem to be doing the same thing. It was worse. You know, what I couldn't ask for more than that. So uh no, and so so I think as a result that's helped to enforce as well the confidence in things like paperless, I hope, is this track record of kind of not making that many errors. And then when there is an error, moving quickly to fix it, of course. Yes.

Robustness, Rollouts, And Trust

SPEAKER_04 37:38

I always tell people when you are working in the IT field and in IT projects or IT in general, you you need to understand that uh a server will be down at some moment in time. There will be bugs in the solution. So you need to be ready, as you said, as you said, what it makes the difference is your capacity to react, how you react to the changes, how you react whenever there are issues. And I think that the level of support we we got from our management in this office as well, from everyone in the office, to manage that transformation is what also, again, I insist uh it makes a difference. Because obviously, if the moment there is a problem, your management is there telling you, you see, this doesn't work, then it is very difficult to have the time to fix the problems or not to very quickly, as you said, roll back the solution and say, okay, we will think more a bit before we roll out a solution. So if you are brief enough, you get the support from management and everyone else to go ahead with these changes, knowing that it will be bumpy. Yes, it could be. Then then then just you can deliver value, as I said.

SPEAKER_02 38:46

So and Hale, I'm gonna do the really annoying thing that I do on the podcast now, and I'm I'm sure I frustrate Grimm because I do this on every podcast. I'm the timekeeper alongside everything else. And whether we're there or thereabouts, I'm afraid, it's it's flown by. We could carry on talking for for a lot longer, I'm sure. And maybe what we do is we think about bringing you back at some point in the future to to explore some more stuff. But are you sat there thinking, oh, I'm glad they never asked me that? Or, oh, I wish I had the chance to say this. Is there anything else you want to add before we before we close down the podcast?

SPEAKER_04 39:17

Well, I usually tell people, um, but it is quite in line with the conversation we have had. I know that a lot of people are very reluctant or very concerned about the use of AI, or they are or they fear about uh the future of the of the office, the future of the role of examiners, the future of the role of formality offices, pattern attorneys. Uh, what I tell people is whether we like it or not, things are going to happen. AI, generative AI, is changing the way we interact with everything we have. Uh from Netflix to Amazon to Google to all products we use on our daily basis, your mobile and everything else. So it is going to happen in our area as well. So better to be ready. Be ready for that, be open to adopt technology, uh, try to get as much support as you can from everyone, including that transformation. It is not an easy change. I I do understand that when you have been doing something in a way for more than 20 or 30 years, if suddenly you need to interact with completely different tools, it is something difficult. But I insist, whether you like it or not, things are going to happen. So better to be ready for that.

SPEAKER_02 40:24

And and also I should put a picture there for SEPA. We have two committees focusing on AI, one in terms of AI and the law and how patent law will change in the future, and one looking at AI and the profession and how the profession will change in the future. So also look to your professional body because there is a lot that they can do to help and support you in the transition to AI. Angel, thank you so much for coming on. It's been lovely having you. Thank you. But you don't get away that lightly, or at least Gillem doesn't get away this lightly. Okay. Because we always have this closing question. Uh Guillem, I've been thinking about this, uh, and it it was the chess references earlier that got me there. So do you remember in your childhood, Gillem, that chess would always be in what we used to call a compendium, a compendium of games, where you'd have like chess and ludo and snakes and ladders and various other games. So I'm simply going to ask you, what is your favourite childhood board game or a game of your choice? What's your finest memory?

SPEAKER_01 41:23

I loved a game called Othello, if you knew of the.

SPEAKER_02 41:25

Oh, yeah, yeah, no, Othello, yeah.

SPEAKER_01 41:27

It's this eight by eight grid, and you have little beads and little pebbles with white and black sides, and then you have to basically put them either side of some of it. Probably a very small vote of go, I'm guessing, or something. But he thought I was really good at it until I started playing computers at kind of basic level and utterly smashed. And I gave up on it. I think it's quite an easy one, probably, for computers to solve. And just quickly on that. I love the fact that my dad was always a big draft sold checkers player. And I remember reading that long time ago, um, they said humans solved drafts. In other words, there was a series of moves that you could do that basically guaranteed to draw. Uh so they had to randomize, they had to randomize the early moves into ornaments just so that you didn't just play the gambit that always won if you started. So I gave up on drafts and basically solved at the point, just find solutions. Yeah, a fellow.

Final Reflections On Adopting AI

SPEAKER_02 42:18

So you gave up on drafts and you ended up drafting.

SPEAKER_01 42:21

Life story.

SPEAKER_02 42:24

And Hell, what about you, sir? Have you got a childhood, fond childhood game that stayed other than chess? Chess playing chess.

SPEAKER_04 42:31

When I look back, I remember spending a lot of time playing with people more than playing alone. I remember more time spent playing football and games with uh my colleagues, uh, going with the bike to different places. At that moment in time, uh I missed those times just because I was spending more time uh without sitting in a table or sitting in front of a computer and more time spending spending more time with people. Um, so no, I I don't remember a part of chess that I was playing that as well because I was playing that at the school. I don't remember um anything concrete. I do remember I have a lot of memories about my time with with my friends and spending very long hours after the school uh doing all kinds of things, but not playing games like that.

Closing And Childhood Games

SPEAKER_02 43:18

So I'm I'm I'm now quite worried because this is portraying too much about our upbringings. So Gillem spoke about kind of fond memories. You've spoken about your fond memories of being kind of outdoors and with friends and so on. And I know Gillam's gonna ask me where I am on this, and I'm now terrified about the answer I'm about to give, and I'm not quite confident in.

SPEAKER_01 43:37

I'm kind of I'm kind of assuming it was something like bear baiting or cock fighting, no me.

SPEAKER_02 43:42

No, no, no, no, no. Could couldn't be further from the truth. So I was in pubs from a very young age, uh, in that my dad would always take me to the pub with him. Yeah. And he would always sit with his friends and play. I I I've never been very good at card games, okay? And it's probably because the first one I ever watched play. Was so complicated that I I thought I would never be able to learn it. And that's cribbage. And they would they would sit and play cribbage, which of course is a card game, but it's also got this to the to the outside world complicated scoring method with actually quite quirky phrases to some. And I won't say them because some of them sound a bit rude, but they're probably not. They've probably got other kind of etymology, but they they they sound a bit rude. But yeah, so cribbage is the one for me. I've never mastered it. My dad was some kind of like local cribbage champion and could peg peg peg round the board, which is what you do, uh at pace. But yeah, no, that was the one for me. I was fascinated by it, but was never any good at it. But yeah, so I spent my time in a pub watching cribbage, my fond memory. Angell, thank you so much for for coming on and and sharing your time and your work and your life with us.

SPEAKER_04 44:44

Thank you to you for the invitation. And yes, very happy to come back whenever we do want to be able to do that. Yeah, yeah, no, we've been discussing.

SPEAKER_02 44:51

We will definitely be inviting you back. Gwillem, I will see you on the next one as ever. And to everybody listening, if you would be so kind as to leave us a little review to say how interesting you found this podcast, and of course the rest of the podcast that we've ever put out, we would be eternally grateful. Thank you all.