3#14 - Claes Lyth Walsø - Towards a Data-Driven Police Force (Nor)

«Dataen i seg selv gir ikke verdi. Hvordan vi bruker den, som er der vi kan hente ut gevinster.» / «Data has no inherent value. How we use it is where we can extract profits.»

Embark on an exploration of what a data-driven Police Force can be, with Claes Lyth Walsø from Politiets IT enhet (The Norwegian Police Forces IT unit).
We explore the profound impact of 'Algo-cracy', where algorithmic governance is no longer a far-off speculation but a tangible reality. Claes, with his wealth of experience transitioning from the private sector to public service, offers unique insights into technology and law enforcement, with the advent of artificial intelligence.

In this episode, we look at the necessity of integrating tech-savvy legal staff into IT organizations, ensuring that the wave of digital transformation respects legal and ethical boundaries and fosters legislative evolution. Our discussion continuous towards siloed data systems and the journey towards improved data sharing. We spotlight the critical role of self-reliant analysis for police officers, probing the tension between technological advancement and the empowerment of individuals on the front lines of law enforcement.

We steer into the transformation that a data-driven culture brings to product development and operational efficiency. The focus is clear: it's not just about crafting cutting-edge solutions but also about fostering their effective utilization and the actionable wisdom they yield. Join us as we recognize the Norwegian Police's place in the technological journey, and the importance of open dialogue in comprehending the transformations reshaping public service and law enforcement.

Here are my key takeaways:

• Norwegian police is working actively to analyse risks and opportunities within new technology and methodology, including how to utilize the potential of AI.
• But any analysis has to happen in the right context, compliant within the boundaries of Norwegian and international law.
• Data Scientists are grouped with Police Officers to ensure domain knowledge is included in the work at any stage.
• Build technological competency, but also ensure the interplay with domain knowledge, police work, and law.
• Juridical and ethical aspects are constantly reviewed and any new solution has to be validated against these boundaries.
• The Norwegian Police is looking for smart and simple solutions with great effect.
• The Norwegian Police is at an exploratory state, intending to understand risk profiles with new technology before utilizing it in service.
• There is a need to stay on top of technological development of the Norwegian Police to ensure law enforcement and the security of the citizens. This cannot be reliant on proprietary technology and services.
• Prioritization and strategic alignment is dependent on top-management involvement.
• Some relevant use cases:
  • Picture recognition (not necessarily face-recognition) - how can we effectively use picture material from e.g. crime scenes or large seizure.
  • Language to text services to e.g. transcribe interrogations and investigations. 
  • Human errors are way harder to quantify and predict then machine errors.
• This is changing towards more cross-functional involvement.
• The IT services is also moving away from project based work, to product based.
• They are also building up a «tech-legal staff», to ensure that legal issues can be discussed as early as possible, consisting of jurists that have technology experience and understanding.
• Data-driven police is much more than just AI:
  • Self-service analysis, even own the line of duty.
  • Providing data ready for consumption.
  • Business intelligence and data insights.
  • Tackling legacy technology, and handling data that is proprietary bound to outdated systems.

Transcript

Data-Driven Politics in Norway

Speaker 1 0:00

This is Metadama, a holistic podcast about data management in Norway. Hi and welcome. My name is Winfried and thank you for joining me on a new episode of the Dama Norges podcast, where we look at how to give data management in Norway a boost, show the competence and at least the level of knowledge we have within Fragfeld, and that is why I invite with me Nordic experts within both data management and information management to a talk. Welcome to today's episode of Metadama. And finally, we have an episode again on Nordic, which I am always very happy about, and I have with me Klaas today. Welcome, hello.

Speaker 1 0:49

It's a very exciting topic that we have in front of us to talk about, and that is data-driven politics. When we talk about data and AI in politics, there are very many who immediately think of minority reports or automated decision-making processes in the police and the judiciary. It's a feeling that you're losing control. So today we're going to talk about where we are actually, what is actually possible and where does the limit go. There are a number of improvements that we have seen in the last few years in different countries, both in the USA, in China, in Israel, where algorithms have been used to make the judiciary more efficient, to make the politics more efficient, but perhaps in some cases it has been good at costing the security of the citizens, and what does this actually mean for the trust we have in politics? It is a term that has been used a lot lately, and that is Algo-crasity, that is, a democracy based on algorithms instead of citizens. Is that what we are on the way to, or are we actually on the way to better policy services, with more effective services, more accessibility to politics and therefore also more security for us as citizens?

Speaker 1 2:04

We are excited to talk about it, but before we dive deep into the topic, mer tilgjengelighet til politiet og dermed og mer sikkerhet for oss som borger Vi er spennende å snakke om, men så har vi dykket dypt inn I tema, så vi er gjerne høre litt mer om Klaas. Så velkommen, klaas. Gjerne presentere deg, jo takk.

Speaker 2 2:16

Klaas Lydvoldse heter jeg. Jeg jobber I det som heter Politiets IT. I have been there for two years. I come originally from the private sector, so before I worked in the police, I worked as a data analyst and technology director in a company called Retriever, which deals with metadata media insight. Before that I worked for 11 years in Telenor A different role, actually, a lot of IT-revelling and the last four years there was also data and insight and personalisation, using data to better understand the customer and do the best for the customer. So now it's less focus on making money but more focus on value for society and the people out there. For the police department, it leads to what we call a delivery area, which is called knowledge. So it's about taking this data and making it into something that you can actually use and understand. And then I have the additional assignment I would say, which is to lead a data-driven policy. Then we can all be responsible for the fact that we are quite far away from extreme cases and also minority reports, but we can go a little further afterwards.

Speaker 1 3:38

We can. What does Klaas do in his free time when you don't think about politics?

Speaker 2 3:42

on your free time, when you don't think about politics. It's not so very exciting, I would say, but it's a bit like a movie. So it's funny that you bring up Minority Report. The film and TV series I read gave me the education to follow the 24 series where it was very, at that time, far-reaching use of technology to get in the right direction and police work. So it's a bit fun to get back to reality, which is of course a bit less exciting. Otherwise it's a bit like family time and training.

Speaker 1 4:17

And it looks like you're living the dream having to go to work.

Speaker 2 4:21

We can say that I think so.

Speaker 1 4:24

And there may also be something there, an original interest for where the AI-invented work comes from. I have a good question.

Speaker 2 4:32

Actually, the first years of my career there was a lot of IT development focus. That is now also. But I took a master's degree in technology management about midway through here and that's where I found fascination with data, how data can be used for new understanding and insight and, not least, how people can use this to change the way they work and create new values for the company you work in. So I did a little training there actually. So I worked in Telenor, as I said. So after I got a fascination there, I was on the sales side and did there, and the role of the retrievers was a bit of planning because they sat in the top management there and would drive the company forward based on data and insight. It was a no-brainer.

Speaker 2 5:33

When the possibility of the police appeared, it was a dream come true. I could finally follow the trail and work in the police, which was a bit of a understating thought. Sporet jobbet I politiet, som har vært litt sånn underliggende tanke. I hvert fall, samtidig som jeg kunne bruke kunnskapen at, det var egentlig en bra match med hva jeg nå kunne bidra med inn I politiet.

Speaker 1 5:53

Veldig spennende. Nå nevnte vi Minority Report to ganger allerede, og jeg tror I fjor sommer så satt vel de fleste nordmenn som harians who are interested in AI and data, sat and read Inga Strømke's book Maskiner som tenker, and she mentions some of the examples we see in other countries when it comes to overworking, when it comes to face recognition. She mentions a startup in Israel. She mentions something about how Chinese authorities treat Muslims in society. She mentions a system that has been in the US for a very long time called Compass.

Speaker 1 6:38

Now I just have to read it Correctional Offender Management Profiling for Alternative Sanctions, which is an automation of which penalties they are given. Objectional offender management profiling for alternative sanctions, som er en automatisering av hvilken straffram og tildel vi har, og så er det noen algoritmer rundt. Hvordan klarer vi å spore terroristiske handlinger før de skjer? For eksempel, så mye av det som vi har sett I Minority Report ser ut som det blir virkelighet, og så er det veldig forståelig at folk blir litt sånn skremt og tenker, and it's understandable that people get scared and think is this what we're on our way to? And I hope you can accept that this is not what we're on our way to. So if you, with your knowledge and insights, are going to be able to explain what a data-driven policy should be. Very much.

Speaker 2 7:29

The examples you gave show a scary picture of what is actually possible today if you let go of this completely. So I'm concerned that we are far from there and we don't want to be there. But what we're doing now, since we're talking a lot about artificial intelligence and machine learning solutions, we also think that a data-driven policy must also understand and master the type of technology or tool it wants. We now have a good group of data scientists who work in the policy department within different areas. It's a lot about understanding the possibilities of technology, but we need to do it in the right context, so that all of those who work with me on this are placed together with political experts, one to find out how this can be used to give good value for the job they do in everyday life. And then we need to take with us the legal and ethical aspects of these things. But this is where it starts to get difficult. Often, when you discuss this in the media or other places, it's extremely difficult. It's very easy to hang on to in and see where it goes. It's very important to find regulations and make sure we go in this direction and secure the rule of law, but we're quite far from there.

Speaker 2 8:54

There are a lot of smart and simple things we can do that give effect to our policy. Today, for example, we look at image analysis, not face recognition, but understanding all the image material that the police have. Then you can think about how you can effectively use it, for example through large reports, instead of people doing it manually. It can be thousands or ten thousand or a hundred thousand of pictures and files for a song. We also look at more like speech to text services. So today you sit and transcribe a good amount of material and, for example, the hearing of children should be be transferred to the full. It can also be automatically reactivated. The point is that we can exploit artificial intelligence in quite safe forms and still create value. It is also about building competence, both on the technical side, but also to get the interaction, which is the difference between technology, technologists, political professionals and lawyers, into what we call double-professional groups. To really understand which problems we should solve and how we can solve them. We always have to make sure that it's well-informed.

Speaker 1 10:15

And there are 10,000 things I want to ask about now. But what I often talk about, which is perhaps something many who are not quite clear about, is that when you talk about text-to-speech and transcription, when you have a person doing the job, there are possibilities for errors, Just like there are possibilities for errors when a machine does it. But the difference is that the error the machine makes you can measure it. You know how much error it is. They say that it's, for example, 10% of the fault margin you have on the machine which you can actually do something with. But when it's human factors, it's much more difficult to measure. It's much more difficult to apply the intention with the errors.

Speaker 2 10:58

That's right, and then you can argue that the machine will do the same mistake, maybe every time again. Maybe, but it will not be solved. But it is important to emphasize that we are still in an unexplored phase. I am very much for or the police. I mean, I need to understand and get to know what the fault margin is. What are the consequences if the machine makes a mistake? What is the consequence if the machine makes a mistake?

Speaker 2 11:27

We know how we can use technology in the right way. For the alternatives, it is put a little simply Either we do nothing but learn that society continues to use artificial intelligence without us making a bigger appeal on how we can use it, or that we end up with okay, but then we can't do this, so we buy a solution from a producer from another country, don't know how the model that is at the bottom is built up or what the film margin is, and so on. So I think it's extremely important and it's the task we have now that we manage to build competence and understand how we should use this in the right way. So the examples I mentioned are still conceptual, very exciting.

Speaker 1 12:12

And then you talked a little about Tvers Havelægget team, and I would like to get a little more understanding of how these teams are set up, because the solutions you build and think about are a point of contact between, just on the one hand, between technology and ethics.

Speaker 2 12:33

That's right To explain a little background. So traditionally, from a few years back, the police department has been a service provider. In the police. We have developed services and orders that are thought through by one of the others and perhaps not so much two-faceted interaction. And as long as you order now, two years later, you get the result, but then it's either too late or it's not what we want, either. Simple, of course.

Tech Legal Staff for Data Management

Speaker 2 13:03

So what we work a lot with in the political industry now is to get away from the project model. We need products that will live over time and to get the good products we need a double-facetedness, both internally in the IT world, so that we have the right type of developers, designers and other roles in the teams, but also the political aspects of the users who are going to use the product, how it is going to be used. And then we have just started to establish what we call a tech legal staff. So we have had a leader since January. We have positions to get more there. The idea is to have lawyers who work for our team. We understand what problems we are trying to solve and, of course, make sure that we are outside the legal landscape, but also maybe take the balls further and get rid of those who are responsible for the handling of the determined data sources, for example. Or maybe take the slightly larger balls further with the directorate and afterwards the Justice Department, because maybe we need law changes to get to what we need to get to. So it's all the interaction we want to get and we work very hard to move there, not only for the art-and-intelligence sector, but also the entire spectrum of IT services in the state. And then, to add a little, we data-driven police are also more than just art-and-intelligence services. We are also more than just a knowledge intelligence service.

Speaker 2 14:43

We also have a lot of focus on getting to what we call self-employed analysis. So it's a lot about working with data when we do it in the right way Make sure that the data flow is in place, with infrastructure and everything that must be there, of course, that we manage to work properly with the data quality and the information management around it, so that the state and the decision-makers whether it's in the directorate and in the board of directors and so on, or if it's a police officer in a police car who needs to make a decision on whether to drive there or there or whatever it is, then it's about making it easier for the data to access and can do the self-reliant analysis in place. In addition, we of course, make a good number of dashboards and finished reports that cover maybe 80% of the needs, so that's a focus area. We've started to get quite good help.

Speaker 2 15:38

Another aspect is that it is no secret that the state has extremely old technology. Some of it is from the early 2000s. It started to become a moment ago, actually 30 years ago. It has also led to the fact that we have linked to the system. First of all, the old information model is completely like that and it is very silo-based, so the data is very connected to the solutions we have, which again leads to the need to move data manually from the system. If you, for example, do research or there has been an event where you collect information, then you have to move that data or put it into a new system in the research phase, and then it's a matter of reporting and so on. So that's also something we look at how we can have data instead of having it right and getting it right internally.

Speaker 2 16:39

Of course, the legal framework in relation to the use of data and the security requirements.

Speaker 2 16:45

There are also other aspects in the multidisciplinary aspect.

Speaker 2 16:49

We have high ambitions with this and we have a lot of focus on data and information management, getting it to fly in the same context as the product teams we want to establish. We often say that it's fun to fly, but it also applies to the implementation management, so that we can really exploit and have quality when we do going to use the data for artificial intelligence and, if I hope, when there is a law change that will allow us to build our own models based on that data, because there is no opening for that today in the police registry law. So that's the limit. I'm not a lawyer, but it's something like that. If we collect data from a different form, for example after research, then there is no hope to use that data to make a model for machinery or coastal intelligence today, but it could be useful because then we get a better feeling about what kind of yes, how franchise-based that model is and if it's possible for us to do it with any issues we can help with in the future.

Speaker 1 18:02

It's a very exciting topic how we can collect data, use it in relation to the object we have collected before, but also ensure that the data we have collected can be shared. We can ensure that we get the insight we need to solve the problems we have. Have you thought a lot about data sharing, access to information from other sources, from other public data, for example?

Data-Driven Decision Making in Law Enforcement

Speaker 2 18:34

We have that, yes, so, first of all, we have already brought in a lot of data today, from register data, for example, the state's VVS, the driver's license, driver's license and other types of data. But, to be honest, we don't share that much and that's probably a bit of a limit on what's interesting or what we can share. Our focus now is to get things in order in our own house before we start sharing. We can share our focus now I get some order in my own house before we start to share. We think so. So on the prize list for 2024, now I really have a fast pace in the professional relationship, which is that. But what we are talking about is to share data internally. So it's a kind of strategic direction that all these teams have that they will make available the data they produce and consume to other teams so that we don't have the duplication of the migration from system to system Again, provided that there are rigorous rules for us to do it again, but now I don't know.

Speaker 2 19:40

This is what lawyers and police professionals have to do. I read, for example, a report after Kongsberg-Hendelsen that was here just a little over two years ago. I have a pill to drink, if you remember. So there are cases where you have collected information in solutions, for example police operations, and then it's not the same solution that you work on on a related side. So I still don't know if it would have been right to share the information, but the reason for it is more the technical. It could well be that it's legal things that hinders us, but it's a bit like we think we have to open up something. We have the whole picture internally in our solutions.

Speaker 1 20:30

It's a very exciting difference to talk about here. You talked about the order of the house. Gustav Ågesen from Lånek has been on the podcast in season one. We talked about data sharing in the public sector and that's a topic he really burns for and is very engaged in. There is a difference between limiting data sharing because it's a legal issue or it's a matter of collection that limits it, or if it is technical solutions that limit it. And in many public data, from what I have seen now, it's been a while since I've been in the public, but you have a very solution approach that is very problem-based. So you find a problem, you find a solution to solve that problem, then a new problem appears, then a new solution comes and that ends up. After a new problem appears, a new solution comes up and that ends up with many different solutions and many different problems that are not managed and the information that is locked in these silos that must be able to be broken up.

Speaker 2 21:33

I agree. So we are also in the process of creating a kind of collaboration arena in the public sector. There will be some initiatives that will see how to get this part of the network to flow across. We are also involved. We have it on radar so we don't wait to do anything. But our focus is mostly on internal cleaning. But of course, the information we have in public bags we have to share and have it one place, Not centralize it, but that it is one place so that we don't get the same information everywhere. At det ligger et sted, så vi slipper å hente den samme informationen alle steder.

Speaker 1 22:21

Hvis vi kan ta et steg tilbake, fordi jeg tror dette er litt spennende å se det I kontekst. Din rolle I politiet er definert rundt kunnskap Og så, helt klassisk, så tenker jeg på den der pyramiden fra data til informasjon, til kunnskap, til wisdom. Er det det som er tanken her? The pyramid from data to information, to knowledge, to wisdom. Is that the idea here? Is the idea that one should add problem solving from a more holistic perspective.

Speaker 2 22:51

Very good question. I think I have to comment a little on the knowledge concept, because it is used very actively in politics. It's a common Norwegian word, of course, but it's used very actively in the police. So you might just want to call it data analysis, but you call it knowledge, that we should be knowledge-driven. So I've been right about it that time when I said data-driven it's not to go back a few steps, but it's a little more common concept in the market.

Speaker 2 23:25

But I think, no matter what you call it, it is necessary to coordinate which problem we want to solve, where we can get the profit, how we can make decisions smartly and get to the Again. This is a multidisciplinary collaboration that is handled because then you of decision making whether the problems will be solved, whether it's decisions on the control room or if it's in the patrol car you have to work closely with that, together with those who understand the data and those who can call it a more analytical team, whether it's artificial intelligence, machine learning or more traditional analysis that needs to be done. So what it's about for me is to get that flow there to work, so that the data itself doesn't matter, it's how you use it, where we can get the profit. So we have to call it knowledge or wisdom, I think. And then we have a direction in politics, for example Sweden, which was just started, but now there is a direction up there.

Speaker 1 24:26

But let's dive a little into the different ways to a data-driven politics, and there are many different areas where you can effectify, different areas where you can use, for example, artificial intelligence in different areas. What are your thoughts on what is most suitable for AI?

Speaker 2 24:50

That's a very good question. We haven't started with our own opinion and we have recruited the expertise. That's my take on it. But what we have now in the last, we have lifted this and established a so-called we call it the owner meeting. The whole time the direction we have set has been anchored in the management and such, but it has perhaps been a little too much driven from the IT side.

Speaker 2 25:19

The purpose of the EIR meeting and getting the top leaders more involved is just about whether we do the right things or not. Is this the right route? I don't know. But then they also have to understand what is actually possible. What competencies do we have? What capabilities do we have? What do we need to do to move on? What is the action room? So it's a discussion we've started working on in parallel. No secret either. We have limited resources, means of money in the public sector, which means that we have to have that dialogue more than we might have had. But when you say that, I know there are a lot of benefits to be gained from where we have started. So the question is is there any other place where there could have been even more benefits to be gained?

Speaker 1 26:17

It's a little exciting, and we have talked about this earlier. The way to accelerate innovation internally is that you have fewer resources to go to and that you have to make it more efficient, that there is a clear and clear need.

Speaker 2 26:30

So I think it's not a bad thing to have less resources when you think about the ability to innovate as a possible direction for how we can really take the big lift and the information models if you can call it that and collect our data in the police in large numbers. We work with small teams to create proof of concept concepts. We also work in the field of weapons to see how it can make life easier for criminal analysts by showing you get more understanding. We have to work on finding the way to the victory. In the end, if we want to show it, we need to work between finding the way to the profit in a big way. It's an exciting thing we're working on right now with very few people.

Speaker 1 27:40

It's very exciting because there are many organizations that are struggling with that. How do I experiment, find new solutions but at the same time have it very close to the operational and quickly usable? Otherwise it often becomes a barrier, because in experimentation you must have the opportunity to come up with new thoughts, you must have the opportunity to fail, take decisions that are not so good, but in the operational you must have a yes or no answer. How are you able to connect these things together?

Speaker 2 28:15

Yes, that was a good question, but it's a little back. If we had just driven on alone, to put it that way, it would have been that we do something from the IT side where it's technically done well, or we guess a little what can be done, and then we get a mix of things. It was interesting, so it's pretty important, I think, all the work we get connected to what are the benefits we can take, what are the problems we try to solve, and then you have to do it together with someone who is sitting and feeling this in everyday life. So otherwise I don't think it's.

Speaker 1 28:50

I think it must be the key to success, unless you are very sharp or have a place, I think and that is the hardest part for us- the policy is a task like many others, which does not only have the part that is very closely linked to the purpose, but also an economy-based business management economy part, which is also important and where there are many opportunities to, for example, use self-employed analysis and optimize the process. How much have you done on this side?

Speaker 2 29:22

Well, I think that's actually what we started with, because it's not just controversial data, it's data we have. We have started to look at our economic data from DFU and our employees data and make solutions to better understand our internal business. People may not know this, but we have 12 political districts in this country and each of them is very autonomous. So, yes, it has its own budget and such, but the administrative law in Napa is quite autonomous. So this is extremely useful, isn't it? How to collect the data that is central to Napa and collect it and give the opportunity to get a much better understanding of how well the district is doing. So that's what we started with. And then the next step is to reduce the level of competence. If we cover 80%, how do you deal with the rest of the 20? Reduce the level of competence so they can do their own analysis or set up their own dashboard via Bopo. And then it's a lot about the cultural.

Speaker 2 30:44

We haven't talked so much about Edva, but it's about working with and constantly exploring the data. That is a major success factor. I think it's a big success factor. It's fast to think that the report shows this and that. So then we do that. Maybe, but that's not 100% certain.

Speaker 2 31:00

You have to ask the right questions and not give them away. I often see people asking me and I say that the police are actually very good at this and this is not what you do in the research. You turn the wheel and knock on people's doors and that's fine. So I try to get a little bit of the underlying idea they have in the same aspect. Here we have a piece of work to do, because I usually say that being data-driven or much of what we've talked about now is much from the technology side. Even if we're not to be professional, there's a lot of technology and solutions, but if you really want to be data driven, it's up to people, culture, processes around it, because if people aren't thinking differently and research and think all the time now we're on many feelings we need data to understand what. When that happens, we won't be able to do much. So it's a difficult part that takes more time than implementing a solution.

Speaker 1 31:59

Yes, it's a difficult part and it's not just politics. It's easy to focus on technology because it's exciting, it's fun and it's perhaps the easiest. I'm not talking much about that, but it's people who are the key to becoming data-driven. What you talked about was data literacy. How can we create a culture where people can read, write, understand and work with data, and also understand what it means to work with data? I thought exactly the same as hva det betyr å jobbe med data. Jeg tenkte akkurat det samme som du sa at folk som har jobbet med etterforskning og analys I politiarbeidet har egentlig en fordel fordi man har gjort det I mange år, Kanskje på et annet nivå, kanskje I en litt mer begrenset dinnærming enn det vi ser for oss nå, Men det er I hvert fall en mulighet å, but it's still a possibility to tap into that mindset. I think so too.

Transformation Through Data-Driven Culture

Speaker 2 32:58

We still have a long way to go before we can really call ourselves a company. In the context I spoke of earlier, it's a lot about getting the right roles in place. We are talking about product development and how we want to change our entire business, or we are tired of doing what we do with the development. So to get the roles in place and to give a fair share of time to the domains that work with data and with information in the daily, to solve these problems and to play further on what the political people already have underlying, I think it can be a good self-reinforcing effect that we want to get people on. But, of course, to turn it around, we can make the best solution of the world if we don't use it or we don't look at the reports or the analyses, or have they been given a little more? Because that's a big question to ask if you think that there are two lines to answer for every time, and that's not the case.

Speaker 1 34:02

It's an exciting journey from what shall we call it the proof-driven to the data-driven. It's probably quite similar to the data-driven part of the show. It's probably pretty close. We're nearing the end of the conversation and I always like to ask the same to all my guests If you have something you want to sum up or a key takeaway, if you want something people can think about in the future, oh, good question, bård.

Speaker 2 34:26

I've already said a lot about culture so I can't go on. But I think it has to be that it's easy to think drastically and extremely, but it can't stop us from thinking the first steps and being part of the journey that happens in the world. So far, reisen som skjer I verden, det blir kanskje ikke så kjempe kanskje, men jeg tror det er det viktigste at vi er med på reisen og har startet det sånn.

Speaker 1 34:58

Og så må jeg gjerne legge til at det, at du stiller på podcast så åpen om disse temaene, synes jeg er veldig bra, og det kan jo føre til at vi kanskje får litt mer god debatt rundt, can lead to a little more good debate about this in Watson. So thank you very much, klaas, you're welcome.