NDE 4.0 Podcast Transcript
Episode 13 —How Artificial Intelligence will transform NDT
Our Guest: Lennart Schulenburg,VisiConsult
Editor’s Note: In the interest of time, this transcript is still in rough format and has not been edited for proper grammar and punctuation. If you have a need for a fully edited transcript, please contact us.
How Artificial Intelligence will transform NDT!
[00:00:00] Lennart Schulenburg: Welcome to floodlight Software’s, NDE 4.0 podcast, where we
[00:00:17] Nasrin Azari: interview various experts in industry 4.0 concepts, issues, and technologies for non-destructive
[00:00:23] Lennart Schulenburg: testing and inspections. This show is the place to go to learn about the biggest challenges
[00:00:28] Nasrin Azari: and opportunities around NDE 4.0 from some of the smartest people
[00:00:32] Lennart Schulenburg: in the industry.
[00:00:34] Lennart Schulenburg: So sit back and be prepared for a really thought provoking discussion. Hope you
[00:00:38] Nasrin Azari: enjoy the episode.
[00:00:51] Nasrin Azari: Hello everyone, and welcome to today’s episode of floodlights N D E 4.0 podcast, where we pose five questions to a variety of [00:01:00] N 4.0 experts and explore the benefits and challenges in this emerging field. Today we are joined by Leonard. Schulenberg, who is an expert in bringing new technologies to the N D T industry in an effort to create safer and more effective inspection systems.
[00:01:17] Nasrin Azari: Leonard Schulenberg is general manager of visit consults, X-ray systems and solutions, which is headquartered in Germany. With over 135 employees globally, the company develops and manufacturer’s high-end x-ray inspection systems for N D T. This includes high precision CT systems, robotic solutions, inline systems, a D R software, and much more.
[00:01:41] Nasrin Azari: I. Leonard studied com computational informatics and graduated through an international M MBA program. As a regular speaker at conferences, he is considered an expert in the field of digitalization and automation of NTT processes. In today’s discussion, we are going to talk about the most impactful technologies [00:02:00] to this industry, automation, robotics, and ai, and I’m super excited to have you with us today.
[00:02:05] Nasrin Azari: Leonard, welcome to the show.
[00:02:07] Lennart Schulenburg: Thank you. Thanks for this kind introduction. I’m looking forward to it.
[00:02:11] Nasrin Azari: Great. Um, so the format of this podcast is that I will pose five questions to Leonard designed to dig into some of the most meaningful and interesting aspects of n E 4.0. Our podcast was developed to help educate and expand conversations around the possibilities, challenges, impacts, and opportunities surrounding N D E 4.0.
[00:02:31] Nasrin Azari: I’m particularly excited about today’s discussion because I know Leonard is a great visionary in the field. So let’s jump right in. I’d like to start with a question to sort of set the stage for our audience. Leonard, can you tell us a little bit about your company VII consult and what your role is specifically at the company?
[00:02:49] Lennart Schulenburg: I. Sure. So our company, as you’ve said, an introduction, we headquarter headquarted in Germany and our main focus is to build, [00:03:00] develop, and install cutting edge x-ray inspection systems for non-destructive testing. So this could be ranging from. Um, manual inspection systems that you take in the field. It could be computation, um, computer tomography systems for doing volume metric, three dimensional inspections or more closer to the industry.
[00:03:21] Lennart Schulenburg: 4.0, um, uh, movement, completely in line. Integrated a automated solutions. So our, our company started in 1996. It was founded by my father back in those days. And our main focus in the beginning has been being a software supplier. So we developed and supplied, um, OEM solutions, uh, so image acquisition, MS processing software, and kind of grew into a full-blown solution provider.
[00:03:50] Lennart Schulenburg: And the funny thing is that we are going full circle because we are going back towards being. More a software driven company as we are talking about artificial intelligence, [00:04:00] cloud connectivity, and digitalization. So my specific role is, uh, being the general manager, taking care about, uh, all the general needs of the company, making sure we have the right strategic fit.
[00:04:13] Lennart Schulenburg: Um, attending, um, like global industry events and making sure the company, um, the company is aligned with the global mega trends. And this brings us, again, back to NDT 4.0 because that’s, from my perspective, the by far, most disruptive and innovative field of n t in the, in the latest years.
[00:04:35] Nasrin Azari: Interesting. So it does seem like the companies come a come a long way and you know, you, you are right on the cut, cutting edge.
[00:04:44] Nasrin Azari: What would you say about, you know, your customers? Would you say that they’re, they’re pretty advanced in digitalization in general. Um, you know, do you find yourself being brought into companies and then having to take a few steps back with them or, you know, [00:05:00] what do you see in terms of the maturity of your customers?
[00:05:05] Lennart Schulenburg: That is all, all over the board. So it is there. There is some industries, like for example, the automotive industry, especially in Europe and in Europe, especially in Germany, which is very, which has a very, very high degree of automation. You see, you, you see, we call these like lights out factories because you could literally switch off the lights.
[00:05:27] Lennart Schulenburg: And the cont, the factory will continue to work. Mm-hmm. Because there’s a lot of robotics, there’s a lot of automation involved, and in these environments, our systems are integrated. Completely into the manufacturing flow, into so-called flow lines. Mm-hmm. And we are, we have, um, a physical connectivity through conveyor systems, robotics, and we have a digital connectivity using like cutting edge interfaces like U P C U A, for example, where you.
[00:05:55] Lennart Schulenburg: Where the whole like data flow and data transmission is, is [00:06:00] automated and that’s what’s called the digital threat. Mm-hmm. That goes through, um, through a facility. So these guys are very, very automated. And then on the other end, this is mainly in very regulated industries like the aerospace industry or also the defense industry.
[00:06:17] Lennart Schulenburg: While you have legacy drawings, legacy inspection requirements, you sometimes have. Really, really Manual processes. Yeah, paper stamps, film, no digital detectors yet. And we are helping those companies to migrate into the digital world and still fulfilling the demanding requirements because that’s always the big, I mean, ultimately, The target of N D T is to provide safety.
[00:06:45] Lennart Schulenburg: Mm-hmm. To make sure we do not have any parts that shouldn’t pass this quality assurance pass. So that’s the, that’s the most important thing, the effectiveness of the quality assurance. But at the same time, we [00:07:00] also have to look at the effectivity and, and that’s where really the digitalization automation really helps companies.
[00:07:07] Lennart Schulenburg: Yeah,
[00:07:08] Nasrin Azari: definitely. Well, really, really good insights. Um, let’s move on to question number two. Which technologies do you believe will have the biggest impact on the N D T industry as a whole? And, and why?
[00:07:21] Lennart Schulenburg: If you, if you ask me for just one, I would say it’s ai, artificial intelligence. Mm-hmm. Yeah. Um, especially in radiography because we are interpreting images and.
[00:07:33] Lennart Schulenburg: AI is just very good at that. It’s very good at pattern recognition. It’s very good at finding features in image data. Mm-hmm. And we, we, we see this already in the medical industry, which is typically two to three years ahead of the entity industry. You see AI solutions beating doctors in radiology. You, you see these big competitions going on, well, like top notch.
[00:07:58] Lennart Schulenburg: Best doctors of the world [00:08:00] are losing to an like losing because that is like in the real world. You wouldn’t say it that way, but in a competition, right? You, you can’t put it that way. Right? And, uh, these, these people are losing in, in, in like interpretation competitions against ai. You see, like for example, the chess computers beating humans in ultra complex tasks.
[00:08:19] Lennart Schulenburg: You start to see self-driving cars. And, and AI really will transform our industry because it will take away many of the repetitive steps and repetitive work that we do, and will replace it through just sheer computational power. Mm-hmm. But the work, this will not put all of us out of work. Our work is just changing.
[00:08:39] Lennart Schulenburg: And that means somebody has to supervise the ai, somebody has to train them, somebody has to qualify them, and that is a complete different skillset as a level two or level three has today. So that is very, very disruptive. And then of course, automation, robotics, cloud connectivity, all that is [00:09:00] has, has their disruptions as well.
[00:09:02] Lennart Schulenburg: And, and that’s why this, this plethora of technology is also bundled into N D T 4.0, uh, which in its, which together is, is ultra disruptive.
[00:09:12] Nasrin Azari: Yeah, really interesting. Yeah. I’m fascinated by the use of of AI and I do know that there’s some fear in the industry around, you know, one of the things that you, you mentioned, which is the, um, people are afraid that.
[00:09:28] Nasrin Azari: AI will, will actually replace people, which in some ways, they, they will, like you said, in a lot of cases people are losing to ais, but I, I almost feel like, you know, combining AI with some of the other technologies like automation, robotics, that is super, super powerful in the industry. It,
[00:09:47] Lennart Schulenburg: it is, and it’s, it, it’s not it.
[00:09:51] Lennart Schulenburg: You can’t stop it. This technology is there and it will be used, and those who use it will use that competitive advantage to either force the others [00:10:00] to use it as well or to go out of business. It’s just the how the economy works. So we can’t just afford to not use a very helpful new tool, but we have to use it smartly.
[00:10:12] Lennart Schulenburg: So we, we need to qualify it very well. We need to train our teams and we need to adapt slowly. I’m, I’m not. And, and this is where I don’t like the word disruption, because it’s not this way that you just say, okay, from today to tomorrow, I change all my processes. I implement an AI and I cut back my level twos by 50%.
[00:10:32] Lennart Schulenburg: That’s what this would be. Um, assure disaster. Mm-hmm. And this is not how, how it’s, it’s more evolution and we need to make sure we have business continuity that we still have. We still can serve our main goal, which is assuring the quality. Yeah. And that should, should always be in our mind as we do the transition.
[00:10:53] Lennart Schulenburg: It’s more a transformation than a radical disruption. And it happens over time. But just to give you an example, [00:11:00] we’ve recently implemented, um, AI tool at a German aerospace, at a big German aerospace company. And we are, we’re the, this AI is inspecting welding lines. And the normal, true positive rate of a human called TPR is roughly 85 to 95 in a at peak percent, and for indications and, um, Our AI out of the box after being parametrized reached, uh, 98.7%.
[00:11:36] Lennart Schulenburg: Mm-hmm. Which is remarkable. Mm-hmm. Especially if you look over fluctuations over the day and as soon as you start to generate all that data and analyze the data, It just becomes so apparent what this technology can do. But at the same time, there are a bunch of level threes and level two sitting, uh, sitting on this data and analyzing the data.
[00:11:58] Lennart Schulenburg: Their job just changed [00:12:00] away from the shop floor into the data space and analyzing generated data. Yeah.
[00:12:07] Nasrin Azari: Yeah. I, I’m totally with you on the, the idea that this is coming and it’s, it’s not worth trying to. To fight against the technology because it’s so valuable. And like you said, your, your, your study is, is one of many that shows how powerful AI can be for the entity industry.
[00:12:26] Nasrin Azari: So, you know, you’ve, you’ve started to jump into this a little bit, but let’s move to question three, which is, what do you believe are the biggest challenges, um, to adopt AI and, and other tech and other ndt, N d E 4.0 technologies in the N D T industry?
[00:12:43] Lennart Schulenburg: Yeah, if you’ve asked me. Two months ago, I would’ve say standardization and standard.
[00:12:48] Lennart Schulenburg: Standard standards. Yeah. But, um, as of recently, a new ASTM standard get got published. That’s, um, that can be used to statistically qualify ai, um, [00:13:00] programs to be used for inspection. So there is now. A guide. It’s not a standard, it’s a guide, an ASTM guide, but that’s close enough that shows how, how companies, even in aerospace environments, because it’s coming out of the aerospace and defense world, can qualify a.
[00:13:19] Lennart Schulenburg: AI is to be used in aerospace, and that’s huge. That’s, that’s the first big step. And as soon as this standard is recognized by the big players in the industry as they see the cost benefits of using that technology or throughput benefits, this will be adopted very, very quickly. But this leaves us with the second biggest issue, and that’s a mindset.
[00:13:42] Lennart Schulenburg: Question. Yeah. That’s, mm-hmm. That’s, um, adoption of new technology is never easy. This technology is so fundamentally different, um, to what has been there before and especially it is many inspectors are. The, [00:14:00] their pride is the, the, the level of quality of their work and that they are incredibly good at spotting, let’s say welding defects, and then this AI comes along.
[00:14:10] Lennart Schulenburg: It’s just better. It’s a stupid computer. And, and, and that, that is a dent to the self-worth of some people that said, Hey, this is, this is my job. This is what I did for 20 years, but, What I, my answer to this is our human brain is much more valuable and it can do much, much more than just finding wealth defects every day, all day long.
[00:14:35] Lennart Schulenburg: Mm-hmm. We can be the masters of the tools that we created and, and AI is really, A tool that got created, and we all read this from time to time, that this general AI that will replace humans, that’s wallock. That that will not gonna happen. But what will happen is that you have very narrow focused, specialized and powerful AI that we can harness.
[00:14:57] Lennart Schulenburg: And as humans, we use those [00:15:00] tools. So what we have to do is a shift of mindset in, in. In how we work, and that is being the masters of those tools and putting them to work and making sure they stay in line and they do what we want them to do. Mm-hmm.
[00:15:16] Nasrin Azari: Yeah. Interesting. It’s, you know, it’s, it’s almost like there’s a, there’s a psychological component to this change process that’s gonna be the biggest resistant factor, it seems, and a lot of folks, when they think about digitalization and, you know, changing digital transformation, changing their business to adopt technologies, I think people rarely think about the sort of the human factor, the psychological factors, which can be the biggest roadblock that they’re gonna have.
[00:15:46] Lennart Schulenburg: It. It is the biggest roadblock of digital transformation, and I don’t talk about ai. I talk about the stuff that you do. Process automation, process digitalization, cloud, cloud, migration, mm-hmm. Is always the hu. It’s always the [00:16:00] human factor. Question. It’s about making sure, first of all, everybody understands what we wanna achieve.
[00:16:05] Lennart Schulenburg: Why are we doing what we are doing? What’s the goal here? Why, why do we even talk about technology? Because AI is cool. No, because AI helps, helps you to improve throughput by 50%. Mm-hmm. That’s why we gonna do it. Mm-hmm. And, and then we need to discuss how we gonna do it. Mm-hmm. How we gonna change the processes?
[00:16:23] Lennart Schulenburg: Um, how we gonna, how, how we gonna manage, uh, to address. Problems. How, how we will address people being threatened, being scared, maybe by technology or just feeling incompetent in the beginning because it’s something new. Mm-hmm. And, and then we have to implement then the what comes, what are we doing to actually address that?
[00:16:42] Lennart Schulenburg: And we need communication. Communication. And if a manager wants to implement these technologies, it’s very, it’s crucial to, to communicate early and communicate good. And yeah, the technology is, Easily solved. I mean, you just contract people like you or [00:17:00] me or others that implement these technologies. They are there, but the big challenge for the companies is to implement them and make sure all employees understand it and follow it.
[00:17:12] Lennart Schulenburg: Yeah. And
[00:17:12] Nasrin Azari: like, and like you said, it basically changes the, the, the whole, um, work, you know, f from our, from my perspective, it really changes how a business operates and people have to get used to that and embrace it. Um, versus sort of that, that fight, that fight against it. Um, So, you know, knowing that there’s quite a bit of resistance for, for a variety of reasons.
[00:17:36] Nasrin Azari: For, for question number four today, what do you think will trigger the tides changing towards more greater acceptance of these new technologies? So, so when do you sort of see that tide changing to the point where it’s obvious that there’s no turning
[00:17:51] Lennart Schulenburg: back? For digitalization in our industry. I think that point has been already reached.
[00:17:59] Lennart Schulenburg: Mm-hmm. Wherever, [00:18:00] wherever it is allowed to digitalize people. Digitalize. And that’s the migration from like digit, from analog film to digital detectors. And that is the pure digitalization step, but that’s just the first foundation layer for all the NDT 4.0 steps. This is kind of like the. Entity, 3.0 step and the NDT 4.0 step is that really builds on it.
[00:18:22] Lennart Schulenburg: And like u u utilizing the data, um, um, using AI or, or driving a automated processes using robotics. And I would say that is happening. And a big driver of that is, is the Covid 19 pandemic. Mm-hmm. Before, especially in aerospace, there has been still a lot of resistance for these technologies. And we were, we were, for, for example, we were propagating a concept since.
[00:18:48] Lennart Schulenburg: Probably two years. We call inspect from home. That means all your systems and all your review stations are connected through a cloud. And this can be a private cloud inside the company. It doesn’t have to be online on Amazon or something, [00:19:00] but it can be completely private and inspectors can log in from wherever and, and, and evaluate the images.
[00:19:07] Lennart Schulenburg: We’ve got so much pushback on that. Mm-hmm. And then this virus comes along and suddenly customers call us, say, we need our people to work from home. Now because we are not allowed to enter the facility. And we’re like, whoa, that’s a 180. Um, two months ago you said this will never happen and now it does.
[00:19:26] Lennart Schulenburg: Mm-hmm. And many people are not switching back. And so what we feel is that Covid is a real, real digital accelerator, and we kind of, we kind of won four years that that would’ve, it would’ve taken probably four years, what has happened in one year. Yeah. And with all this new adoption, people see. Oh, it’s not that bad.
[00:19:46] Lennart Schulenburg: It, it wasn’t that hard. It actually worked. And people get a little bit more curious and a little bit more excited to try something new, and that’s what we see. Mm-hmm. So I, I definitely see a lot of traction and a lot of [00:20:00] movement on many of those technology fields. Yeah.
[00:20:03] Nasrin Azari: Yeah. I think, um, It’s really interesting how people don’t think that they can, they can, you know, they, they can make a change like this and then when they’re forced to do it, it’s amazing.
[00:20:15] Nasrin Azari: Um, Not only to see what they can accomplish when they thought that they couldn’t with when they thought they couldn’t. But then like you said, it, it starts to change people’s perceptions on, you know, what’s possible. And the amazing thing is how quickly it all happened, you know? And we’ve got some similar stories where we’ve talked with folks that.
[00:20:37] Nasrin Azari: Were literally inspection companies that were literally shut down during the pandemic because they didn’t have any digital tools and they literally could not work remotely. And then there were some that had started putting, putting, um, some capabilities in place and they were much less affected by the pandemic.
[00:20:55] Nasrin Azari: So it’s kind of funny to think about how. The pandemic has affected our [00:21:00] businesses and our industry like has and sort of accelerated it. So there’s really, that’s a good story. You know, one of the good stories around sort of this, you know, the, the negative thing of the pandemic certainly has, has its good points as we’ve seen.
[00:21:15] Lennart Schulenburg: Um, yeah, that’s true. There’s, yeah, where, where there’s light, there’s shadow, and where there’s shadow, there’s light. And, um, there were definitely, uh, good aspects. Um, Among the many bad aspects, obviously. Yeah.
[00:21:28] Nasrin Azari: Yeah. Um, and I do think that, that as we move forward, you know, there are continue going to continue to be some, some roadblocks that stand in the way.
[00:21:38] Nasrin Azari: What do you see, um, in terms of, of roadblocks that we might stumble over as we try to adopt these technologies going forward?
[00:21:47] Lennart Schulenburg: Yeah. Um, I, I don’t see so many technical roadblocks, at least in our industry because this stuff is really, really coming together. Um, okay. As, as far as I can say. But what, what is, [00:22:00] and this I think we also had in our discussion panel discussion on the ass and t conference, there’s a lot of over-promising, um, there’s a lot of misuse of buzzwords, there’s a lot of expectation around.
[00:22:13] Lennart Schulenburg: And there’s also a lot of people misre, misrepresenting the technology either for promotion or because they don’t know what they talk about. Yeah. And that is very dangerous because, um, you know, you know the, the hype cycle, right? Mm-hmm. And, and many, many of these buzzwords really get overhyped and blown out of proportion.
[00:22:33] Lennart Schulenburg: Everybody goes crazy. And then comes the. The reality, reality of peers. You crash and you, and everybody says, okay, I stop. All investments management was exciting. And then no cost savings come. And then they stop the program. But it’s more healthy if we just have a realistic, very realistic view of what can be achieved, what we can do and what, and also what we cannot, what other limitations and just be honest about it.[00:23:00]
[00:23:00] Lennart Schulenburg: And that unfortunately, I miss, especially with some solution providers on the market. It’s just, oh, ai, ai, ai is coming. It will solve everything. No, it won’t. Mm-hmm. It, it just won’t. Um, but it will help you in these areas, but it probably won’t help you in those. Yeah. And so that is one of the speed bumps I might see coming, because T 4.0 is just, It, it just went big so quick and suddenly everybody talked about it.
[00:23:30] Lennart Schulenburg: Yeah. But not everything was really NT 4.0 something was just marketing labeled things that we’ve still, that everybody has been doing for 10 years. And that’s just taking away the credibility of the movement. Right.
[00:23:44] Nasrin Azari: And, and you know, even though ai, you know, you were talking about the, um, you know, the, the.
[00:23:53] Nasrin Azari: The rates or the, the percentage of, of, um, positive, you know, detection and [00:24:00] your AI might be at 97% and your, in your person might be at, let’s say 92%. So obviously the AI is better, but it’s, it’s, I feel like there’s a little bit of a, you know, a. A bias in that if an AI makes an error, like one single error, that could be, like, that’s gonna be a really big deal, right?
[00:24:20] Nasrin Azari: Oh gosh, the AI made a mistake. Whereas people make mistakes all the time. And if you say, okay, this person hit 92%, you know, positive or correct, that’s a great thing. But if an AI makes one mistake out of a hundred, then I, I feel like there’s this bias towards. You know, kind of over exaggerating that quote, failure.
[00:24:43] Lennart Schulenburg: It’s, it’s a very good point. I’m not a big Tesla fan or something, but it’s always, you know, this autopilot. There’s one crash or there’s one person being hit by a self-driving car, that’s a tragedy. Sure. But you have, I don’t know how many cock crashes a day and [00:25:00] nobody talks about it. And you’re absolutely right.
[00:25:02] Lennart Schulenburg: It’s just the newness of technology that’s, that’s, it’s, it’s, it’s in the spotlight. And Yeah, but you, you need to measure it by the same rules and KPIs as you do with normal processes. And that’s what the new guide, for example, that the A S D M developed does. They are actually doing double blind studies and comparing operators to AI and nobody knows who’s who.
[00:25:25] Lennart Schulenburg: And, and, uh, actually applying the same math and the same statistics to whoever made the decision, because ultimately it doesn’t ma matter where the decision comes from. And that’s how we have to see it. We have to take the emotions out. But, but this is unfortunately, if, if, if people are hesitant, if they’re afraid, that’s when the emotions come in and that’s when over there it made a mistake.
[00:25:47] Lennart Schulenburg: Yeah. Yeah. It doesn’t work. I told you so,
[00:25:49] Nasrin Azari: like Right, right. Yeah, definitely. Um, definitely. So inter it’s gonna be very interesting as this unfolds over the years. Um, but I, I. It’s [00:26:00] really interesting to hear all the different studies and, and how we might move forward in a, in a positive way that’s maybe a little bit, not quite as hyped up as, as we think it might be.
[00:26:09] Nasrin Azari: Um, so let’s move on to our final question today, which is, you know, for those people and organizations that are interested, In, um, you know, moving forward with AI or some advanced, uh, N D E 4.0 technologies, how can you know, what do you recommend in terms of people that want to start their journey?
[00:26:31] Lennart Schulenburg: I. Well, don’t underestimate the journey.
[00:26:35] Lennart Schulenburg: First of all, it’s not like I buy an AI and I use it. No, you have to train the ai. You have to adapt the AI to your part. You have to, you have to, um, implement your quality requirements. Um, there might be some fine tuning necessary. Also, and I don’t say this because we are a solution provider, you can go to anybody else, but don’t try it on your own.
[00:26:57] Lennart Schulenburg: It, it absolutely is a very, very long [00:27:00] progress. And it took us many, many projects until we got to a like model that that really does the job. And we trained thousands of images, um, into this ai. And that’s also the biggest. Issue with ai, it needs a lot of data. Not not, this is why it’s called big data. You need a lot of labeled data, and that’s very time consuming.
[00:27:22] Lennart Schulenburg: And we have a team of, I think eight people who are doing nothing different than labeling indications and training the models, optimizing the models, and making sure the AI is improving. And the more data you have, the better the network gets. And that’s also why you will not see. Nobody would come, would, would get the idea to, to create a new Google because it, it’s pointless because there are so much data in their search engine that it doesn’t make sense to build another one.
[00:27:52] Lennart Schulenburg: It, it just would, you would not beat the performance. It would take way, way, way too long. And it’s the same with ai. If you would start to [00:28:00] build your own AI from scratch, Well hire 10 or 15 people and do this for two or three years, and you might have a good solution. Mm-hmm. But, but this is typically not the case.
[00:28:11] Lennart Schulenburg: So my, my suggestion is reach out to people that are doing this already and, and talk to them. Mm-hmm. And again, I’m not saying this as self-marketing, but reach out to anybody and there’s a couple of solution providers in that field, and they will, they will show you the journey towards, towards an AI system.
[00:28:29] Lennart Schulenburg: So what
[00:28:30] Nasrin Azari: do you think about companies collaborating with their data? That’s something I’ve touched on a little bit with some of the other guests we’ve had in this podcast, but I feel like there’s a resistance, um, that companies are very protective of their data and I feel like there is a really, really good.
[00:28:49] Nasrin Azari: You know, potential for, for, um, you know, sort of, sort of amazing, um, benefit that can be gained by companies [00:29:00] that are willing to sort of pool their data together. What do you think about that? Do you have any experience in working with your customers about, you know, kind of sharing their data for the common good in order to sort of train AI and, you know, be part of maybe a, a, a global solution?
[00:29:18] Lennart Schulenburg: The interesting, I, I hear this a lot and I did not experience it. So first of all, okay, it’s not like you have to share your data. It’s. Data is being labeled and it’s being used as input for a model, and then it goes away that there is no re residual of that data in your model. It’s not like you would open that AI and look at all the data of other companies.
[00:29:42] Lennart Schulenburg: It’s literally gone because what, what, what, what has done is the neuronal structure, like in our brain, formed by what the AI saw. It’s like a. Kid learning in a new language. Mm-hmm. And, and the synapses in, in, in its head are, are firing. [00:30:00] And then the, the, the, the book is gone, but you can read whatever other, um, other text in that language you can see.
[00:30:08] Lennart Schulenburg: And the same with ai. So the models are trained. And, and, and, and, and get smarter by it. So it’s, it’s not really a question of data sharing here, because that data is a disposable one-time thing that has to be shown to the AI to, to get better, so. Mm-hmm. I don’t think we will from my per, I mean, at least with our customers, we don’t have that discussion a lot because we can guarantee every customer, nobody else sees your data.
[00:30:40] Lennart Schulenburg: Mm-hmm. But Right. Okay. But then the question we ask is, do you wanna benefit from other data the AI has seen? If the question, if the answer is yes, others will also benefit from the data that the AI has seen from you. Right. So that’s kinda what I was a choice. Mm-hmm. [00:31:00] Interesting. And, and if they say, no, I don’t want that, that means.
[00:31:06] Lennart Schulenburg: It gets a lot more costly because then we will have to start from scratch for that company. Well, not 100% from scratch, but we’ll have to start with a much, much bigger data set and cost, triple, quadruple. And this is how we really, at one point, ai, AI decisions will become very, very cheap if every company develops or gets developed, their AI on their own.
[00:31:30] Lennart Schulenburg: This is, uh, we talk about probably 200, 300,000 euro of investments. For each company, but at from, from the point on, we have developed a model that has been fine tuned. So many times we are talking about about sense per per decision. Only because, because the AI can be adapted to whatever image they see at one point.
[00:31:56] Lennart Schulenburg: It’s like you don’t train your brain to read just one piece of [00:32:00] paper, but you learn the language and read whatever you want. Right? And that’s the, that’s what we have to achieve to bring down the costs of decision and to reach a real level of effectiveness in artificial intelligence in our industry.
[00:32:13] Lennart Schulenburg: Yeah. Really,
[00:32:14] Nasrin Azari: really interesting. Well, um, this was an amazing discussion today. Do you have any last thoughts before we, um, before we close today?
[00:32:24] Lennart Schulenburg: No, you had some great questions. You, you, you kind of, uh, facilitated all, all my important points. Um, you know what I, I think the most important is to all the listeners, be realistic about what the technology can do.
[00:32:38] Lennart Schulenburg: Don’t overhype it, uh, but also be open and, um, don’t be afraid of it, but also be realistic. And be respectful because, This, the things that are coming will change our industry. Yes, they, they will render some jobs obsolete. They will change the structure of how things are done. [00:33:00] And that will happen whether we like it or not.
[00:33:02] Lennart Schulenburg: Uh, it is just inevitable. It will come. The only thing that we can do is be proactive and react before it happens to us and be the, be the people that, that, that steer the change and be the, be the, the people that use the tools and not be the ones, um, yeah, go going in collision with them. Yeah.
[00:33:23] Nasrin Azari: Yes, great points.
[00:33:24] Nasrin Azari: Thank you so much, Leonard, for being here today and bringing us such great insights. I really look forward to the interesting advancements that VII consult will bring to the N D T industry and other organizations as well. Um, I encourage our listeners to follow Leonard and VII consult. You can find links on our podcast webpage.
[00:33:44] Nasrin Azari: And if any of our listeners have feedback or would like to nominate an individual or an organization to be a guest on a future episode, please send a message to one of us here. Do the contact us form on our website, www.floodlightsoft.com. [00:34:00] Thanks again for joining us and see you next time.
[00:34:03] Lennart Schulenburg: Thank you.
[00:34:16] Lennart Schulenburg: To
[00:34:16] Nasrin Azari: learn more about NDE 4.0 emerging technologies and digital transformation, please visit
[00:34:23] Lennart Schulenburg: www.floodlightsoft.com for additional resources, including our blog and several
[00:34:28] Nasrin Azari: relevant white papers. If you have any
[00:34:31] Lennart Schulenburg: questions about today’s episode or suggestions for future episodes, please send an email to info floodlight soft.com.
[00:34:39] Lennart Schulenburg: Thank you so much.
For more expert views on NDE 4.0, subscribe to the Floodlight Software blog at floodlightsoft.com.