We’ve detected that you are using an outdated browser. This will prevent you from accessing certain features. Update browser

Risk, ‘psychic numbing’, and wellbeing take centre stage at Lloyd’s Register Foundation International Conference 2019

The Global Safety Podcast, Episode 5 - The Safety Accelerator

Each year an estimated 2.8 million people die from accidents in the workplace or from work-related illnesses and a significantly higher number are injured. The subsequent fall-out can be very damaging. Reputations can be ruined overnight, share prices crash and consumer confidence tumbles.

So in this edition of the Global Safety Podcast, Tom Heap explores the world of Safetytech: technology being applied to safety solutions. It’s a sector with huge opportunities for growth. It’s estimated that the potential market in developing and creating new ‘safetytech’ is worth somewhere in the region of $863 billion by 2023!

This edition of The Global Safety Podcast features insights from:

  • Jan Pryzdatek, Director of Technologies, Lloyd’s Register Foundation
  • Kyle DuPont, CEO, Ohalo
  • Stuart Feffer, Co-founder and CEO of Reality AI

 

Listen now wherever you get your podcasts:

 

Listen on Apple Podcasts

Listen on Spotify

 

 

Listen on Google Podcasts

Listen on Stitcher

 

The Global Safety Podcast investigates the biggest safety issues facing the planet and looks at the latest science and innovations being developed to safeguard our future in an unpredictable world.

Episode transcript

Tom Heap [00:00:11] We usually take our health and safety at work for granted and hope to live our lives free from harm. Yet it's estimated that more than 2.8 million people die each year from accidents in the workplace or from work related illnesses. And a significant higher number, again, are injured.

[00:00:29] Incredibly, work related accidents are estimated to cost three point nine percent of global GDP annually, and much of that costs hits the individual companies involved. And the fallout can be very damaging. In other ways, reputations can be ruined overnight. Share prices crash. Consumer confidence just vanishes. So could technology play a greater role in protecting the lives of workers and the survival of businesses? Welcome to Safety Tech Safety Tech. Is technology being applied to safety solutions? And there are so many opportunities to be had. It's estimated that the potential market in developing and creating new safety tech is worth somewhere in the region of 860 billion dollars by 2023. And this is where Lloyd's Register Foundation comes in with their safety accelerator, an open innovation system that allows startups to develop their own innovations and bring them to the market.

[00:01:28] I'm Tom Heap and welcome to the Global Safety Podcast. Once again, we're on Zoom, but that doesn't mean we can't highlight the technologies under development with the accelerator program.

[00:01:39] Much more on that later. But now I want to introduce who's with me, experts, innovators. Let's meet them first. I'm joined by Jan Przydatek, director of technologies at Lloyd's Register Foundation and an expert on all things relating to the accelerator. And I'm also joined by two innovators championing their own safety tech breakthrough. Kyle Dupont is CEO of Harlow and Stuart Pfeffer is co-founder and CEO of Reality AI. And we won't keep you waiting too long to reveal what their innovation actually is. So to get things underway, a quickfire question for each of you, what do you think is the biggest safety concern? And we're talking wide here that you'd like to be able to solve. And you never know. We might come up with some cracking ideas on this episode. Jan, what do you think's the real challenge out there?

 

Jan Przydatek [00:02:31] Accidents often happen because we don't see them coming because we don't think they all happen to us. Sometimes this because we simply don't know that the risk is there and sometimes it's because we do things the way we've always done them. Or maybe we want to take shortcuts because that makes our lives easier. And often we don't understand how the actions we take contribute to something altogether, much bigger. So if we could solve these points step by step, I think the number of accidents that you were talking about earlier would probably be reduced.

 

Tom Heap [00:03:03] And is there a kind of technological way to getting us to realize those perils rather than just assume everything will be OK?

 

Jan Przydatek [00:03:11] It's not only technology. Technology has a role to play, but this is about people as well. Technology can help people to make better decisions. But actually, it's that's the people that are the core and the way that this will actually change. It's the people stupid.

 

Tom Heap [00:03:27] What's your biggest safety concern?

 

Stuart Pfeffer [00:03:28] Well, I think you answered it very well. Accidents happen because we don't see them coming and we take that literally. That is from an engineer's point of view. Seeing it coming reduces to a sensing challenge, a reality. That's what we focus on, is using advanced sensing to see things coming, even if it's not the visual. That's really what we do. We provide advanced engineering tools for using sensors to figure out when something is happening that shouldn't be happening or predict the possibility of something happening.

 

[00:04:04] That that that's how that's how we view it.

 

Tom Heap [00:04:06] You almost want to be inventing the safety version of the praecox from Minority Report that could see see a bad thing about to happen in a way.

 

Stuart Pfeffer [00:04:15] And, you know, we CCS because as human beings, we often we we immediately go to the visual. But it often means hear or feel the vibration or notice something strange in an electrical signal. But one way or another, finding the things in the environment that allow someone to detect that something bad is happening or could happen is the first step.

 

Tom Heap [00:04:40] Feel a vibration. We've gone straight from Minority Report to Spiders and Spider Man, but let's move. Might leave pop culture aside for a moment. Kyle Dupont, what's the biggest challenge out there for you?

 

Kyle DuPont [00:04:49] Mean, I think probably build on both Yun's and Stuart's comments. Safety incidents are recorded often in just unstructured text. It's just text that somebody saves on a computer somewhere. And this text is being saved not only within a single organization, but at organizations throughout the UK and throughout the world. These safety incidents aren't necessarily unique to those organizations. And a lot of learning can be had when you use the knowledge from those incidents across multiple organizations. Now, what we find is a big problem, and that is that there's a lot of private data in there and we need to be able to unleash that data while also respecting privacy as regards to medical conditions, as regards to sensitive sites, et cetera. And so what we do at a hollow is essentially help bring those hundreds of thousands, if not millions of documents together in a privacy respecting way so that that data can be shared between organizations.

 

Tom Heap [00:05:42] Will come on to some of the details of your particular innovations in a moment. But, John, if I could just come back to you, how would you sort of define the umbrella of safety tech? And why is it so important for Lloyds Register Foundation?

 

Jan Przydatek [00:05:54] Well, safety tech is all about the technologies we can bring together to impact safety. And by that, what I mean is about insight. One of the things that we need to be able to understand is what's happening so that we can take appropriate steps to to stop things going wrong. And what safety tech does is it uses safety technology solutions to make products and services that we can use in the real world to tell us what's happening so that we can do something about it. Safety accelerators, essentially, it's a bridge between companies that understand the safety problem. They own the safety problem and technology start ups that have solutions for those problems start by working with the problem owners to define exactly what their problem is. And we call them our challenge partners in this process. And then through our partner plug and play in the States. It's a company that is a venture capital company in itself, but also it's a large effectively an accelerator company. It helps industry find technical companies that together can work on solutions. Together, we find startups that may have solutions to these problems and it helps to incubate them and teach them how to do their business better.

 

Tom Heap [00:07:14] Can I just interrupt for a moment when you talk about the problem owners, are you talking about, you know, the big companies, be they Unilever or BP or, you know, those those kind of they could be engineering, they could be service companies. It could be anything. They all have risks involved.

 

Jan Przydatek [00:07:32] They could I mean, it's the companies that make the things and products and services that we all buy and rely on. So you've got people like BP and other energy companies, transportation companies, through to the food industry as well. They all have potential for unsafe things to happen.

 

Tom Heap [00:07:49] I guess all this might be seen as a little bit of a cost up front for business, but I guess nothing compared to the financial impact of a serious accident. Can you put some figures on those little accidents?

 

Jan Przydatek [00:08:01] Accidents can cost an awful lot of money. If you look at somebody like the BP accident,.

 

Tom Heap [00:08:06] You're talking about Deepwater Horizon.

 

Jan Przydatek [00:08:07] I mean, the Gulf oil spill. Yes.

 

Jan Przydatek [00:08:09] Yes. But in general, fatalities in companies can affect their share price. You can also lose the ability to to manufacture. You may lose the ability to do what you do. And the costs of that are something you have to pick up later on.

 

Tom Heap [00:08:24] I've seen a figure that the fatality or serious injury and additional loss can be over over 100 million for a company at least.

 

Jan Przydatek [00:08:32] That's right. And these these are serious amounts of money for any organization. And the point is that they with the right insights, they can be avoided.

 

Tom Heap [00:08:41] Well, let's talk about some of the particular tech and just briefly Jan to stick with you for a moment that one of those on the accelerator is this senseye technology. Can you tell me what that is, how it works?

 

Jan Przydatek [00:08:52] Okay, so it's a great example of technology being used for safety and being transferable across sectors. And that's really important with the safety tech that the companies that are in the safety tech business are not necessarily starting out as companies looking at safety. They're working in other sectors and then moving that technology for a purpose which is different. And Senseye is exactly one of those organizations. When you talk to David, the CEO, he talks about things like look into my eyes like a magician, I can do so. And people have often said that the eyes of the window of the soul and it can tell you anything. People have looked into this for a long time without success. But the secret of Senseye is that it understands that there are 5000 muscle fibers in the eye which are connected to different parts of the brain. And if you know what you're looking for with the sensing technology that captures the images of your eye, you're able to tell how somebody is. You can tell if they're intoxicated with alcohol or on drugs. You can tell if they're tired. You can tell something about their stress, their mental condition. And if you have this information in real time, it means you understand the people that are going about to do a particular job and in the safety accelerate. So this was applied on the bridge of a ship where when the crew were coming on board for duty, they could be tested to see if they were fit for duty or for the duties that they were particularly supposed to be doing that day. If you're able to understand people's condition before they go on duty, you're very much able to prevent accidents from happening.

 

Tom Heap [00:10:35] And is it more sophisticated than just I mean, up to a point, you and I can look at someone's face and say, well, they look either really tired or maybe they've had a few. It's obviously being much more precise than that.

 

Jan Przydatek [00:10:46] It is. So when it's looking into your eyes, it's very much able to pick up on the things that are happening to you. But the one thing that we've learned is that people have different tolerances. So you need to be able to understand someone's tolerances. So what shows up as one person having not had a good night's sleep may impact somebody else quite differently. So it's an indicator. It gives you an insight of things, but a decision will always rest with a person should they be able to work or not?

 

Tom Heap [00:11:16] Well, that's senseye, which sounds really good. But let's move on to are innovators who are in the virtual room. So let's start with you, Kyle Dupont. Tell me about your company and what it is that you've invented.

 

Kyle DuPont [00:11:29] So at Ohala, we sell a product called the Data X Ray, and the data X really comes from the data privacy space, being able to identify what data is private and then being able to extract that to a useful format. And then lastly, redact that. Now, what we did with Lloyds Register and the Health and Safety Executive, which is the UK's health and safety regulator, was take that that last bit, essentially the redaction bit and display that at scale. And the health and safety executive essentially had 600000 documents of safety incidents. Right. That they wanted to analyze so that they could hopefully, you know, at the end of the day, find out where accidents are happening and get people home safely to their families at the end of the day. Now, after they did manual redaction. For about 2000 of those documents, they went to Lloyd's register and said, you know, help us, and we were fortunate to be selected by the plug and play program to help out. Let's register on redacting those documents.

 

Tom Heap [00:12:26] So let me just pause you there, Kyle, to make sure I've understood it as we're going along. So these are the kind of reports that come in from every safety related incident across the country. You're always supposed to file an accident or near Miss Report and things like that. That's right. So they actually have got all these documents and they want to be able to learn from them. But there's data protection issues there. Is that right? Yeah, exactly.

 

Kyle DuPont [00:12:49] So so it has things like, you know, very simple stuff like Kyle Dupont fell off a ladder and broke his neck and can't walk anymore.

 

Kyle DuPont [00:12:56] And that's sensitive personal information that shouldn't be shared with third parties. Right. HSBC been entrusted to hold that data and they're not supposed to send that to third parties without proper protections around that data. Now, you can just go through it with a with a black marker and pen and just say, you know, let's take out the Kyle Dupont bit and then we'll be able to say that that data is properly protected. But after they got through around 2000 of those documents, they found that it was going to take them up to 12 and a half person years to do that, which, you know, is a pretty much a nonstarter. Right, for for kind of unleashing that data. So when they when they went to register, we were able to do the same thing with Lloyd's register and about a day of machine time instead of two and a half person years of time, you know, reducing their costs. I think it was like nine acts and reducing the time by, I think four thousand five hundred X or something like that. So it was quite a quite a good achievement, I think. And now since then, we're we're actually working on them with some extra other projects.

 

Tom Heap [00:13:52] I'm really fascinated by this. So how can your technology I mean, some some of this may be commercially sensitive, so I'll let you police that boundary. But but how can your technology look into the documents so that you can see documents? That's fine. And what it understands, what names are on signs, what addresses are it understands what other sensitive information is it taught me through how we can do that.

 

Kyle DuPont [00:14:14] Yeah, well, I mean, it's kind of like having, you know, hundreds of thousands of analysts at your beck and call a machine can read hundreds of thousands of words per second and you can scale it up and down as much as you like. Whereas you and I maybe if we're fast, we read ten seconds or so.

 

[00:14:29] So it's essentially the ability to get that scale of human readability and similar accuracy rates, but within a machine so that you can scale it up and down. And you don't have to keep these people on board all the time to do it.

 

[00:14:42] And you can save the analyst time for the stuff that they actually, you know, are good at, which is analyzing the data for accidents,.

 

Tom Heap [00:14:48] But in a very simple term, can look at a document and see the name Tom Heap or Kyle Dupont and think that isn't an English word as I know it. And it's got a capital letter. So it could well be someone's name basically. Is that kind of thing?

 

Kyle DuPont [00:15:03] Yes, it is. So so we have a pretty sophisticated analytical pipeline, which probably isn't so consumable on a podcast type of thing, but it uses various techniques, that is natural language processing, more traditional techniques like dictionary matching regular expressions. And we mix all that up together to build unique algorithms that can look through free text and find the sensitive bits, essentially.

 

Tom Heap [00:15:27] And this may be very obvious to you, but what is the big win for the idea of getting hold of this big data analysis?

 

Kyle DuPont [00:15:33] Well, the first step is just simply being able to, you know, analyze the data in a way that respects privacy. Right. The next step is essentially looking across those reports and finding, you know, typologies of accidents like our ladder accidents often associated with inclement weather. Or are there certain type of accidents on particular site types, like if you're if you're at the power plant or a certain type of accident, more more likely there. And that's really where you want to get to, is is being able to find out where those types of accidents are happening so that people that are performing those type of activities can can take care not to not to have those kind of accidents.

 

[00:16:10] Eventually where you want to get to is is more of a cross organizational data sharing platform where, you know, you can have people ingesting data from all over the place and being able to use that data at scale to find out how accidents are happening across the world and being able to apply that to their own local state.

 

Tom Heap [00:16:28] And presumably you can cross-reference things are saying there's a lot of accidents happening around 4:00 p.m. I don't know exactly detail people's blood and that matches with bad weather or a certain type of behavior or precise. I guess those are things. And that gets you to where we started.

 

[00:16:45] That gets you close to the precog thing that we talked about, being able to predict when and how an action might happen and shutting that door.

 

Kyle DuPont [00:16:54] Right, right. Right. And that's exactly right.

 

Tom Heap [00:16:56] That's a really nice use of our machine learning. I often hear the terms. I'm a bit kind of skeptical about what they're actually delivering, but I really do get that. Well, let's move on. Stuart Pfeffer, tell me about your company and what it is that you're delivering.

 

Stuart Pfeffer [00:17:10] Sure. So the reality to a little bit different. We providing, on the one hand, Adriaan tools for engineers who build stuff with sensors and on the other hand, we will sometimes use those tools ourselves and build solutions, which then are offered to the market. In the case of the Lloyd's Register Accelerator, Lloyd's has paired up with Sellafield Ltd. in the National Nuclear Laboratory to try and solve a challenge involving duct work in nuclear power plants.

 

[00:17:44] There's quite a lot of duct work in these plants and they carry various gases from place to place.

 

[00:17:49] And when there is a problem in the duct work, so that that's a problem. So the challenge is to try and figure out problems while they're still very small and easily repairable. So and things like holes, corrosion, loose connections. And we are now working within an hour in Sellafield to use sensors to try and spot them.

 

Tom Heap [00:18:12] And just before you come to your solution, tell me how that, you know, rattling bolt or, Rusty Patch would have been seen in the past or even today, I guess. Yeah.

 

Stuart Pfeffer [00:18:23] You know, I'm not sure I even know we're not nuclear engineers. We're, you know, I people and signal processing engineers. But I presume they have mechanisms for figuring out where pressure is dropping or there's visual inspections or the problem is manifested in some other larger way.

 

Tom Heap [00:18:41] But what is the weakness in the current system that yours could address?

 

Stuart Pfeffer [00:18:46] Well, catching it very, very early. Right.

 

Stuart Pfeffer [00:18:49] So the notion here is that by putting a sensor in a transducer, every certain amount of distance along these rather long pieces of ductwork, we can tell that the sound of the steam moving through that duct work is different. So we listen for the sound of the steam. We listen to the sound at the joints. Every now and then we issue a little ping and see how the return of that ping is different if it's different than it was when it was normal. We know that something isn't normal and you can pinpoint it.

 

Tom Heap [00:19:26] So a lot of these pipes have sort of certain properties, resonances, maybe other other things. I don't know, conductivity, heat transfer, various things that you're kind of sensing.

 

Stuart Pfeffer [00:19:39] Yeah, I'm sure there's lots of those things because it's inexpensive. We're going to be using we're using sound. We're at the very early stages with Sellafield and in ah on that particular solution. But yeah, the sound and vibration has the advantage with current currently available commercial components are being very, very inexpensive to deliver.

 

Tom Heap [00:19:59] And I guess certainly in the Sellafield context, there's a sort of double safety gain here. Presumably the actual inspection system, the human inspection system can be quite hazardous, certainly if you're involving anything to do with potential radioactivity and then you've got the massive safety thing of it involving some kind of accident.

 

Stuart Pfeffer [00:20:16] Yeah, you know, I think they're looking for broader coverage at lower cost and higher reliability. You know, the thing about using a system like this is that you can watch the entire system all of the time as opposed to periodic inspection regimen.

 

Tom Heap [00:20:33] And this is a technology that presumably could be applied to other areas of big engineering, you know, refineries, for instance, that also have a lot of ductwork or steel or cement plants or anything. Absolutely.

 

Stuart Pfeffer [00:20:45] We have another customer. This runs chemical plants in Japan and they're, you know, using our stuff to detect filter clogs and, you know, that type of thing. But, you know, on the safety category, we've also, you know, just released an automotive solution, this one future mobility award a couple of weeks ago, where we're also using sound to hear the sound of something coming around a blind corner, go on. Commercial collision avoidance systems and cars are largely relying on cameras, radar, ultrasound, line of sight technologies. They can only see what they can see. But if you've ever driven with the windows down, you know that sometimes if you can't see it, you can hear it. Yeah, yeah. So we've put microphones on the outside of a car and turned the car into a rolling microphone array, working with partners like Infineon, the semiconductor company and Denso and Hagiwara in Japan, the car parts companies and built a prototype version of a supplementary system for those camera based collision avoidance things that we'll hear something coming and let you know even if you can't see it.

 

Tom Heap [00:21:55] That's really nice. Briefly coming back to the pipe and sound and residents seeing the ducting. Just going to try a little analogy on you that might help me to understand it might be completely wrong, but famously, a cracked bell makes a different sound from a sound bell hobel without a crack. Is that. The same technology that things have a kind of resonance and if they've got some kind of fault in them, they sound different if you strike them.

 

Stuart Pfeffer [00:22:20] That is exactly what it is. And we can in some cases, do it with passive sound. That is just the you know, as things are moving through the pipes, they sound a certain way. And when they move through damaged pipes, they sound a different way. And we can do it with active sound where we put out a ping and the way the ping propagates is different. So both of those are useful.

 

Tom Heap [00:22:44] Famously, the reason the Big Ben sounds the way it does the last bomb is because it's cracked a bit flat to go down. So I'm interested to hear from all three of you briefly. Start with you, Carl. I mean, where do you start with something like this? Did you come up with the problem or is as a bit of I think John mentioned earlier, sometimes there's a solution out there in search of a problem. How did how did it work for you?

 

Kyle DuPont [00:23:09] Well, our bread and butter is really the data privacy space. So Hollo came out of really GDP and the other privacy regulations that are sweeping the world right now, helping companies with the first two steps that I talked to identify and extract. So being able to identify what type of data is sensitive and then be able to extract that to useful workflows around data, subject access requests and things like that, where we were really able to leverage the accelerator was was on that last step and really developing out that redact pipeline where where, you know, we're going a step beyond that and actually processing the data so that it's usable in a privacy respecting way. So that's really where that kind of evolution came into place.

 

Tom Heap [00:23:47] So you're already trying to make the awkwardness of GDP, which we all talk about, sort of rub those corners off and make it more user friendly within our current world. That was already the space you were in.

 

Kyle DuPont [00:23:59] Yeah, exactly. I mean, the regulations come into place. What do you do? You call your lawyer and lawyers create some processes and those processes are often really hard to actually implement and the real world.

 

[00:24:09] So, you know, we're helping them implement that at scale, at the actual data level.

 

Tom Heap [00:24:13] And so the safety application of this is one of many applications of what your companies.

 

Kyle DuPont [00:24:19] Yeah, we do. I mean, we do. Yes, safety. We do data minimization strategies where we're helping companies delete debt at scale.

 

[00:24:26] We're helping them map that data data, subject access, request response, that kind of thing as well.

 

Tom Heap [00:24:31] And Stuart, same question to you. Did you sort of come to this deliberately or did the safety element of this sort of sneak up on you?

 

Stuart Pfeffer [00:24:36] Oh, no, we were definitely a hammer looking for nails. You know, our stuff was actually originally built for military and intelligence applications. It's no accident that we were in the Maryland suburbs of Washington, D.C. that were the Department of Defense. Is that what you're telling me? Perimeter defenses in Virginia. But I mean, this is where the soldiers and spies do their thing. So that's where this technology originally comes from. And a few years ago, my co-founder, Jeff Saraki, he's the mathematician who actually invented this stuff. And I took this stuff out of a company that was doing it for the military and created a reality HR, which is 100 percent commercially focused. We don't have any government or military, no military customers anyway.

 

Tom Heap [00:25:20] Both Kyle and Stewart really loving those two technologies, because what I love about this is they started as quite impenetrable to me, but I've absolutely got them. And I hope that the excitement of that translates to the audience as well. But I just want to broaden out a little bit in the safety space.

 

Tom Heap [00:25:34] Yeah, I see that overall statistics suggest that the great improvements, the steep improvements we've had in workplace safety are slightly plateauing and industries not getting safer in the way it was in in previous decades. Why do you think that is?

 

Jan Przydatek [00:25:50] I think perhaps it's worth qualifying that statement a little bit, because if you look further back across many decades, then safety has been getting better. We've had regulations, safety equipment, safety processes in place. But what we're finding is that in developed economies in particular, that the rates of improvement has plateaued. And in some cases, even if you look at the United States, there are indications it's starting to trend in the wrong direction. And also, if you look at other economies that are developing rapidly, there are indications that they're also tending towards this plateau. So it's an interesting thing. And the reasons for this could be any one of those is that the way we look at safety traditionally is by looking at the past. We look at historical things that have happened or gone wrong, perhaps accidents, and we collect data from the past and and try to infer something from it from quite a long time ago. And these sorts of things mean that you're not working with the real time actual data that can tell you when something is happening. So one reason we think things have stopped getting better is because we're looking at things in a in a in a historical sort of way. But things are also getting more complicated. The systems. We have the design. Perhaps some time ago, they have barriers in place and processes in place that are there to prevent accidents from happening, but things can change over time. And it's those changes that we might not be picking up so much. We may not even know that we're making changes to impact something that was previously thought as important.

 

Tom Heap [00:27:33] A couple of brief questions. I won't let Kyle come in because I see him nodding, but just that that thing in the States that it could be going the wrong way. I mean, how come?

 

Jan Przydatek [00:27:44] Well, I think it goes back to these things. It's that it's the lack of information of what's happening and what's happening today in real time. If you're relying on the past, it's not necessarily an indicator of the future.

 

Tom Heap [00:27:56] I don't want to sound like the accidents friend here, but there is, of course, another possibility not not in an uptick, but in a plateauing, is it? It becomes more and more difficult to get rid of the last accidents. You know, this is the sort of case in anything, isn't it? To make something 90 percent safe is is great. And I'm not saying it's easy. You know what I mean? It's doable. To get rid of that last 10 percent is always going to be really, really tough. I mean, it could just be an element of that. And this conundrum, it's possible.

 

Jan Przydatek [00:28:22] But I think from what we see will be look at accidents that have happened, some of them will certainly be preventable if there was more information available and if that information was turned into insight that you could do something with and whether that's looking at previous records like we've got Kyle's work, they're looking at how could we crunch lots of data sets of accidents that have happened, but also looking forward to be able to sense things that are happening in real time like you heard from Stuart, and be able to see what normal looks like. And if you understand what normal looks like, then you start to see what things look like when they move away from normal.

 

Kyle DuPont [00:29:00] Yeah, I mean, I think it's just a kind of a general trend and any kind of industry or technology. Right. I guess you have different modalities that you can use as history kind of progresses. Right. So, you know, if you have the ability to implement a regulation that requires certain equipment, that's that's pretty low hanging fruit. And then then you have the ability to analyze massive quantities of data, that's something even better. And then if you have the ability to analyze massive quantities of data across multiple companies because you're able to respect privacy, that's that's another thing altogether. And I think this is all just dependent upon general, I guess, advances in technology. Right. We're kind of at a golden age and natural language processing right now. And we're able to actually get intelligence out of massive quantities of data without requiring, you know, hundreds of thousands of analysts to do it.

 

Tom Heap [00:29:45] How committed to this are big companies? Because I look back at, you know, big disasters all the way back, you know, the Bhopal agrochemical accident in India, which is kind of seared in the memory Deepwater Horizon. We mentioned a moment ago, Exxon Valdez. I mean, these are things that have massive impact on companies, reputational damage, share price, et cetera. So are they still really keen to improve their safety?

 

Jan Przydatek [00:30:09] Let's face it, nobody cares about business intending to do harm, and all companies want to do things that are safe for everyone that's involved. So, yes, what we're finding is that the companies that we're talking see through the accelerator are both big and small. I think there is a real interest from all scales of organization to tackle the safety challenges that they see.

 

Kyle DuPont [00:30:34] I'd like to come in on that one as well. Obviously, this is one of the first forays that we had into the safety tax base. But since then we've had probably four or five other companies come to us with the same problem. You know, these safety teams have the ear of the board and a lot of these large industrial companies are. And so I've been pretty pleased at how willing these companies are to pick up this tech and move with it.

 

Tom Heap [00:30:57] Yeah, and this may relate to your areas here, but I've got a note here saying the Internet of Things is responsible for 77 percent of the safety tech market, which is ends up like many hundreds of billions of dollars. Figure when people talk about the Internet of things in this space, are they talking about work like yours, Kyle? Is that right?

 

Kyle DuPont [00:31:14] Well, I think Internet of Things would probably be more on the Stewart side of Stewart. Wanted to take that one.

 

Stuart Pfeffer [00:31:18] Yeah, I think that that's that's more, more, more our stuff.

 

Stuart Pfeffer [00:31:20] I mean, certainly Internet of Things refers to the ability to deploy lots of devices and have them communicate in real time and constantly. You know, we think of Internet of Things as being mostly about glue. That is, it's the ability to take a device in one place and have it report what it sees to something in the center. You know, if Viotti is the glue edge, which is what we talk about is sort of the brains, that is to be able to put that sensing smarts, deploy it on an inexpensive chip and put that into a device that you can make hundreds or thousands of and deploy along the duct works in a nuclear power plant.

 

Tom Heap [00:32:03] So would you see other ways in which the Internet of Things, the advances that it brings, could help with safety in a much broader sphere of business? Well, certainly. I mean, I think.

 

Stuart Pfeffer [00:32:13] You've already you've already mentioned that you can see industrial applications and not just in ductwork, but in all kinds of industrial equipment. We have customers who are making gas turbines. We have customers who are doing all sorts of industrial processes where there are conditions which may be difficult to detect, where putting more sensing technology and more smarts closer to the location of the source of the problem can help you find it quicker and cheaper. So, yeah, we're seeing that in many, many sectors.

 

Tom Heap [00:32:53] You can just just give me the kind of core pitch of how the Lloyd's Register Foundation safety accelerator can help those who are working on a Problem-Solving technology.

 

Jan Przydatek [00:33:01] So the safety accelerator starts with the people that own the problem actually starts with understanding who has a problem and exactly what it is. And then what it does is it engages with the startup tech community to see who has a solution to that particular problem. And very often it's it's a solution that has never been looked at in safety, but actually it fits in really well. And once we go through a competitive process of picking the technology that we want to try out in the real world with with our challenge partner, we then found a pilot where that is tested to see how it actually performs in real world conditions. And Kyle, for example, is the HSBC testing it on their data. And we've got Stewart testing this technology on nuclear installation or a mock up of a nuclear installation. So it's it's about bringing these two different sides together and creating something that's amazing.

 

Tom Heap [00:33:59] And have you been running us for a few years or is this is early days or what?

 

Jan Przydatek [00:34:02] Yeah. So it's been running now for coming up to two and a half years and it's had quite a lot of successes along the way.

 

Jan Przydatek [00:34:10] So we're very proud of what it's been able to achieve so far.

 

Tom Heap [00:34:12] And presumably planning on running it out for the foreseeable future to find those real safety breakthrough companies.

 

Jan Przydatek [00:34:19] Our aim is that if we see that the potential for safety set for safety tech is sort of an industry that's worth eight hundred and sixty three billion dollars from our perspective, if the world could be made eight hundred and sixty three billion dollars safer, then that would be an amazing thing.

 

Tom Heap [00:34:36] How close can we get to eliminating all risk from the workplace?

 

Jan Przydatek [00:34:39] I don't think we can eliminate all risk, but I think we've got huge improvements that can be made.

 

Tom Heap [00:34:44] Do you agree with that? Kyle and Stewart?

 

Stuart Pfeffer [00:34:46] For sure we're working on it every day.

 

Kyle DuPont [00:34:48] Yeah, I mean, it depends on the workplace you're talking about, right? I mean, sometimes it might make sense to replace a man with a machine or sometimes it might be possible to reduce incidents to, you know, once in a blue moon. Right. I think, you know, there's always going to be people in the mix at some places. So, you know, that means that there's going to be incidents that happened. But I'm a bit more optimistic. I think I think we can even, you know, reduce accidents completely in some activities.

 

Tom Heap [00:35:14] Very good. Or an optimistic note to end on and generally a very upbeat conversation. I was really, really inspired by so much of that. Thank you all for joining me on this edition of the Global Safety Podcast, brought to you by Lloyds Register Foundation. Thanks very much to Jann Przydatek from Lloyds Register Foundation and also Kyle Dupont.

 

Tom Heap [00:35:34] Thanks very much. And Stuart Pfeffer, thank you. Thanks for joining. The Global Safety Podcast brought to you by Lloyd's Register Foundation, please subscribe Zannini, don't miss an episode.

 

latest news

Can't find what you are looking for?

Hit enter or the arrow to search Hit enter to search

Search icon

Are you looking for?