Podcast: How AI Is Transforming The Aftermarket

Don't miss a single episode. Subscribe to Aviation Week's MRO Podcast in Apple Podcasts, Spotify or wherever you get podcasts.

Discover all of our podcasts on our Apple Podcasts channel or at aviationweek.com/podcasts.

Transcript

James Pozzi:

Welcome to the MRO podcast. I'm James Pozzi, MRO editor for the EMEA region. And today, we are exploring the topic of artificial intelligence and its use in the aftermarket and the wider MRO supply chain.

The general public likely knows of artificial intelligence through programs such as ChatGPT, but the entire technological concept is being rolled out across many industries worldwide through different tools and functions. Aviation, of course, is among the industries doing this, examining artificial intelligence. And we'll see today maybe some of the challenges and opportunities that this technology presents.

So, joining me today to discuss this are two Aviation Week editors who've covered this topic in some detail over the past few years. Lindsay Bjerregaard, who is Aviation Week's managing editor for MRO, and Michael Bruno, who is Aviation Week's executive editor for business. Lindsay, Michael, thank you for joining us today.

Lindsay Bjerregaard:

Thanks, James.

Michael Bruno:

Thank you.

James Pozzi:

So, just to start, then, we'll start with Lindsay. You've written extensively about this topic in recent times, including a feature written, I think it was published in the October 2023 edition of Inside MRO, which has since been nominated for an Aerospace Media Award. So, congratulations on that.

Lindsay Bjerregaard:

Thank you.

James Pozzi:

But yes, maybe let's just get a broader perspective on AI in the aftermarket right now. In terms of the technology, where are we at, at the moment, in terms of the main players? I'm thinking OEMs, MROs, and airlines, how are they exploring this technology?

Lindsay Bjerregaard:

Sure. So, there are a ton of different potential use cases that I've seen so far. So, I might not even be scratching the surface, but some of the main ones I've seen, you mentioned ChatGPT in your intro, which is generative AI, and I've seen some startups that are using AI technology in that way for MRO. So, it's kind of like a ChatGPT assistant for technicians. So, they could use an AI app to find answers to what-if scenarios, troubleshoot issues based on data that's coming from things like maintenance logs or sensors.

Speaking of tools that a technician might be able to use, I've seen companies using natural language processing, which is another different type of AI, to instantly log and transcribe maintenance info via speech, which is helpful for technicians to be able to do things hands-free. They could use that kind of technology to more quickly and easily find info that they need, potentially things like finding part numbers more quickly and that sort of thing.

One use case that I think some companies were working on well before this became such a hyped topic, was building simulation models to determine when different types of assets or equipment are deviating from expected or normal behavior to help with optimization of maintenance schedules. And then, one that we're seeing particularly a lot on the engine side is using AI to help detect defects more easily. So, things like automated image processing and assessment for aircraft inspections and engine borescope inspections. And the thought behind that is that AI can be used to help improve the reliability, the speed, the consistency of these inspections compared to human perception.

And since that article that you mentioned, it seems like almost every OEM and airline that we talk to is at least tentatively looking into AI, if not actively working on testing applications. So, earlier this year at MRO Americas, we actually had a panel on AI, and these were some examples that we heard from that. Delta TechOps is using AI to forecast when parts will need to be scrapped. They were doing testing on Pratt & Whitney PW2000 engines. United Airlines mentioned one case that they're using generative AI for is to generate messaging around things like delays on the customer side. Boeing, unsurprisingly, has developed their own large language model using maintenance records, maintenance logs, reports, things like that.

Speaking of OEMs, both GE Aerospace and Rolls-Royce have both recently partnered with some startups on AI. So, GE is working with a company called Waygate Technologies on AI-powered engine blade inspections. They also said that they're using AI in Singapore to improve fluorescent penetrant inspections. Rolls-Royce recently signed a deal with an AI company called Aerogility, to utilize AI for insights into things like asset life cycles and carbon footprint. And then, MHIRJ, which is the maintenance arm of Mitsubishi Regional Jets, they're looking at how AI could be used for some of these previously discussed applications. Things like planning for spare parts and repairs, giving technicians access to AI at their fingertips, and then some of those customer-facing applications like technical help desk and that sort of thing. So, there's a lot of action happening there.

James Pozzi:

Yeah, absolutely. In a short space of time, that's a very broad overview and gives a real insight into the different, I guess, players in the industry that are using, or at least, exploring AI. So, there's a lot of activity happening there.

Bringing in Michael, what are you seeing in terms of AI being rolled out perhaps across the wider aerospace supply chain, and how is that being used? What are you seeing there?

Michael Bruno:

Well, first of all, I just want to pile on and encourage everybody listening to go read Lindsay's article from last October. It's called AI Fever Sweeps MRO, But Will The Excitement Last? And it happens to be the best darn article I've read about AI in aerospace and defense, including at our competitors. So, everybody should go check it out because I think it's a great, great piece that'll give you 10 times as much detail as what Lindsay just told you.

AI, it's kind of interesting because, and Lindsay mentioned this, but AI itself, it's been around quite a bit obviously, and when you talk to aerospace people, they say, "Well, AI has been in our flight decks. AI is the autopilot." So, we think of AI in aerospace as having been around for quite a long, long time, but there are two big things that are happening right now, which is why we're talking about it. One of which is there's a new version of AI, this generative AI, large language models, like Lindsay said. ChatGPT is the most recognized brand around it. And it's been a step function increase in the ability of this technology to do this automatic syncing. And I'm not a computer scientist, but from everything I've read and everybody's told me, I mean, it's a real step function increase. It's here and it's only going to grow.

The other thing is that now you are starting to see applications of this new technology across the board in aerospace and defense. It's everywhere from how a company operates, and Lindsay talked a lot about that in the MRO space, whether you're trying to allow a technician and a hangar to more quickly access documents and scroll through them to find what he or she needs to know, to the chief financial officer back at the headquarters and how they're going to use AI to do everything from complete their books quicker, to supply chain management and try to get ahead using this kind of watch tower idea about how can you start to really predict where your supply chain may go wrong. And they don't even know it, but AI is helping you learn and figure out things before they can even predict it themselves.

To, finally, potentially actually helping engineers and other frontline folks, maybe even pilots someday, but helping folks like engineers who design, whether it's whole aircraft or subsystems, or just parts, to be able to do that more efficiently and faster. And that's actually one of the most unique things, I think, is developing this year. And we're starting to hear about it at some companies, where they're using this generative AI to assist the engineer types to be able to design something, or at least, not waste so much time at a junior engineer level by trying to reinvent the wheel.

They're able to turn to the AI technology and say, "Would you help me come up with a really crude wheel?" And then, I'm going to hammer that a little bit more and I'm going to go to my senior engineer, which I would have done anyway, but now the senior engineer is going to have the knowledge that I went through things probably better and faster than I could have individually, with the help of this AI. And it's applications like that, that we're starting to hear wider across the industry, which is why we're hearing about it now.

James Pozzi:

Yeah. And in terms of the usage, it sounds to me, from someone who's still getting to grips with the topic of AI, that we're very much at the start of the journey. I know you said it's been around for a long time, but I'm talking about in terms of industry usage and all these applications that are developing. Is that a fair comment, that we're at the start of the journey and, I guess, we're only seeing the tip of the iceberg right now, it's just going to continue to grow and be a dominant force? Right?

Michael Bruno:

I absolutely do think this, and I want to caveat my statement by saying, I'm generally a skeptic when it comes to new technologies. We've heard about how additive manufacturing was going to upend metal bending in this sector, and we've heard about big data and the cloud and cyber everything. There was a point there we were putting cyber on the front of, "Hi, I'm cyber Michael Bruno." We were calling cyber everything. And they have, not that they were complete fakes, but they didn't amount to needle-moving technologies. And we are now already seeing needle-moving capabilities emerge from generative AI. It is real, it is here.

And what convinces me of this is how wide the application is. And I just mentioned that, we're seeing it everywhere from an FBO to the back office, to the engineering department, to the CFO's office. Everybody's making use of this. And this is maybe part and parcel as to why it's only going to grow. It's coming at a time of changing demographics in the industry. And it is going to directly help, and I will say, hurt also, and maybe we'll talk about this in a moment, but it's going to initially very much help the demographic changes, particularly the lack of workers and lack of skilled and experienced workers that we are now suffering through in this kind of bathtub in aerospace and defense. But there is a downside to that, and we can talk about that in a moment.

James Pozzi:

Yeah, absolutely. We'll get onto the workforce question shortly, because I think it's very important to do so in the context of this. But back to Lindsay, your take, I guess, on the timeframe of AI and, I guess, at what stage of the journey we're on now. I mean, you've spoken to a lot of companies in recent times about the topic, and did you get any indications of where they see it heading and what speed it's going to accelerate?

Lindsay Bjerregaard:

Sure. So, I mean, I think similar to what Michael said, right? Some of the more established companies like a Boeing or some company that's been using predictive maintenance within their aircraft and that sort of thing, they've already been using the technology, technically, for a while now. However, a lot of these new applications, generative AI and that sort of thing, I think a lot of companies are just now starting to look into that. So, I think we're at the infancy of that stage of it, and it'll remain to be seen, right? I guess, the full scope of how this is going to end up being used.

And there are obviously a lot of concerns about the technology, and one of them being, can we trust that this data is accurate? Right? One of the main things you hear about AI with any industry is, can we trust the answers it's giving us? What kind of data pool is this using? Is it having an AI hallucination? That sort of thing. And particularly in an industry like aviation where we're concerned about safety, it's highly regulated. I think the ability for some of these companies to actually be able to deploy this technology and have the regulators approve it, is going to take time. Particularly, the FAA is always criticized for being slow to approve new technologies. I think it'll be a while before we see companies really using this without having a human backup checking everything.

Speaking of a human backup, one of the things that we've been talking about is the concerns or the benefits for workforce. Right? And so, Michael alluded to this, the MRO industry already has a big workforce shortage. It's got a juniority problem, as Boeing has said. There's a lot of technicians entering the industry now who only have five years or less of experience. And then, a lot of these experienced workers are retiring. So, they're going to have to figure out how to do more with fewer people. And I think this is one of the areas in which they could potentially leverage technology to do that.

I've heard a lot of companies talk about using AI to help these more inexperienced technicians benefit from this institutional knowledge that's being lost. A technician with experience could be talking to one of these ChatGPT-like tools to say, "Hey, when we encounter an issue like this, very often it's this kind of thing." You don't have to go hunt this down in a manual. We can ask the application, that's the AI application, "Hey, we're seeing these things. What could this possibly be?" And it might point you in the right direction. A lot of AI companies mentioned that AI is not looking to replace people, it's more focused on these kind of dumb, repetitive tasks. And so, if the AI can do that, then it frees up technicians for more important work. So, I think that's a big part of that as well.

One thing too, if we're talking about accuracy of results for something like these inspection applications. There was a study out of the University of Canterbury in New Zealand that was testing engine MRO inspections compared to human operators. And it did show that AI was generally better than human operators at the inspection speed, consistency and reduction in false positive rates. However, it wasn't better than humans at everything. So, I think there's always going to have to be a person involved at some point, to review the data and that sort of thing.

And like Michael, I'm particularly skeptical about new technologies, and AI in particular. I think I might be sensitive about this as a journalist where people keep talking about they want to replace us with AI. Right? However, I know for a fact that some of the editors at Aviation Week, myself included, and journalists, use AI tools for things like natural language processing to transcribe our interviews. It saves a lot of time. It's something where it's like we go back and we'll review it, to make sure that it's accurate, but it saves us a ton of time. And so, I can absolutely see there being applications like that, where it could be used. However, I don't think it's a catchall.

And actually, it was kind of funny, Michael, you mentioned CFOs a couple of times, and there was a quote that I loved from that article back in October. I was speaking to the MRO lead for Cranfield University's Digital Aviation Research and Technology Center. His name is Ip-Shing Fan. Very smart guy. And he mentioned that AI can be an attractive label to talk to the CFO and persuade them to give me money, but he very much stressed that it needs to be implemented in ways that are realistic, actually helpful and responsible. So, I think that's something that we really need to keep in mind as well.

James Pozzi:

And you can find a link to Lindsay's artificial intelligence article from Inside MRO, in the show notes.

Bringing Michael back in, actually, I think it poses the long-term question, I mean, it's so obvious that AI does have immediate benefits, and you both explained those in really great detail on this podcast. But what I'm trying to ask, I guess is, when is there too much AI? Maybe working into how it would impact on the workforce long term and this great fear about replacing humans with automated technologies, for example. When does AI become too much? Is there a danger?

Michael Bruno:

Yeah, it is a danger. And I don't really think there's any way around that. Yeah. Let me start with my personal biases. I believe that old adage about technology is neither good nor evil, it's the application of it, which is entirely up to us humans who have created the technology in the first place. So, AI is no different in any other way.

And to the degree that it can in a capitalistic society, we will take this new technology and try to run with it as fast and hard as we can to save costs. And particularly when you have this worker shortage, that is here to stay in aerospace and defense, and it's not just in MRO, it's across the board. We're seeing it everywhere from flight decks to even companies and small tier-three manufacturers who can't get parts out the door to build new aircraft, because they literally just don't have the shifts. They can barely staff a second shift. So, if you could somehow use this technology to address all of that, that's great, but that's a short to medium-term application with a clear benefit.

When you run this out longer, and longer probably means five years or so, so by the end of this decade, you will see companies employing the AI specifically to probably cut their workforces at that point, because long story short, the cycle of new aircraft will have mostly run, we'll be at a different point where companies are looking to lean out their operations even more, to try to squeeze those last few dollars. This is all natural. This is the way free market society works and all that. I'm not trying to call it bad, but you’ve got to look at where the technology is coming into the business cycle. And realizing that we're going to try to use this eventually to not just make up the gap we have, we're going to try to use it to lower costs even further. And the chief way you're going to do that is through people, eliminating jobs, and closing facilities. And it's never been any different in the history of business, let alone the history of aerospace.

So, when do you get to that point? We can debate. My personal belief is by the end of this decade, we're certainly seeing that happen. And I feel like that's the downside, clearly, when we get to a point where technology is literally helping companies lower the size of their workforces, because to me, and I know the two of you also, but we all love aerospace, right? And it's a little hard to love aerospace when you're the only one in the room. You want to share this with other people. We want the workforces to grow. We want more people with their hands on aircraft, because it's so joyful. And as we continue to see the overall aerospace and defense workforce decline from millions, four or five millions, decades ago, to one to two million now, and you think it's only going to get smaller, it's depressing. But it's also, I don't think that's great in the end for the industry, because you're just winnowing down the number of people who care.

James Pozzi:

Absolutely. Yeah. And that's some interesting thoughts to close on, a lot to consider on this topic, and it's definitely something we'll come back to a later date, I'm sure, because it seems like there's going to be a lot of developments in quite a short space of time, and it's certainly going to have an impact. So, not just on MRO, but the whole aerospace and aviation industry. So, definitely something we'll get back to for another day.

But we've come to the end of the podcast. So, thank you very much again to Lindsay and Michael for joining us today. And don't miss the next episode by subscribing to the MRO podcast wherever you listen to your podcasts. And one last request, if you're listening in Apple Podcasts and want to support this podcast, please leave us a star rating or write a review. Thank you.