EMS@C-LEVEL
As Forbes, Entrepreneur, Fast Company and SCOOP writer, Philip Stoten, continues to talk to EMS (Electronic Manufacturing Services) executives he learns more about their individual and collective experiences and their expectations for their own businesses and for the entire electronic manufacturing industry.
EMS@C-LEVEL
Revolutionizing Everything, Even Manufacturing: The Transformative Power of AI with Emily Laird
Use Left/Right to seek, Home/End to jump to start or end. Hold shift to jump forward or backward.
What if AI could revolutionize the manufacturing industry by making data not just accessible, but actionable? Join us as we uncover the transformative power of AI with Emily Laird, an AI integration technologist and fellow podcaster. Emily offers a wealth of experience, sharing her insights on how generative AI is reshaping manufacturing by enhancing data connectivity and delivering predictive insights. We explore the exciting challenges and opportunities in making historical data useful, improving predictive maintenance, and refining decision-making processes. Emily highlights the potential of AI to predict demand cycles and connect consumer signals throughout the supply chain, drawing fascinating parallels with the fast fashion industry. Discover how AI's adaptability offers innovative solutions to design optimization and material efficiency, accelerating industry processes like never before.
As we navigate the impact of AI on job roles, we tackle the ever-present concern of AI replacing human jobs, adding a touch of humor to lighten the mood. AI isn't here to take over; it's here to assist, enabling individuals to specialize by automating tasks outside their primary roles. We dive into the balance between AI and human expertise, focusing on maintaining tribal knowledge and training new employees. With customized AI training and multilingual systems, personalized learning and efficient knowledge sharing become possible. Emily and I express optimism about AI’s evolving role, painting a vivid picture of a future where AI and humans work together to enhance productivity and innovation.
Check out Emily's podcast - Generative AI 101 at https://podcasts.apple.com/gb/podcast/generative-ai-101/id1750985562 or wherever you get your podcasts.
You can also reach Emily on LinkedIn at https://www.linkedin.com/in/meet-emily-laird/
EMS@C-Level is hosted by global inspection leaders Koh Young (https://www.kohyoung.com) and Global Electronics Association (https://www.electronics.org)
You can see video versions of all of the EMS@C-Level pods on our YouTube playlist.
Hello, from my home to yours, welcome to EMS at Sea Level. I am joined by Emily Led, a fellow podcaster. Emily, thanks for joining me today.
Speaker 2Thank you so much for having me, Philip. I really appreciate it.
Speaker 1We are fresh out of the election. We can talk a bit about the use of AI in elections at some point, but yeah, long day yesterday. So nice to for the industry at least and most industries nice to at least have a decision made and that in the review mirror so people can plan forward. I wanted to start just by getting a brief introduction to you and kind of what led you to the Generative AI 101 podcast.
Speaker 2Absolutely Great question. So my background I am an AI integration technologist at the university that I work for, and then I also am a lecturer here that I teach foundational AI, and then I also teach generative AI for writing and communication. I also work with cancer research as the vice president of a business development space, and then I oh gosh, there's so much and then I also do consulting in industry spaces, helping folks bring generative AI into their spaces or making connections at the university I work for to bring in AI and automation into their spaces.
Speaker 1So it's a quite diverse range.
Speaker 1Yeah, and, as I mentioned, those manufacturing episodes are what really drew me in, but I've stayed for everything else. It did exactly the job and I really like that. They're these short punchy podcasts. There's, you know, kind of a bit of your view on what's going on in that particular sector. The finance ones were interesting from that point of view because it felt like there was a personal connection there. There was a personal connection there when you look specifically at manufacturing and in your work that you're doing in consultancy, have you had the opportunity to do much in the manufacturing space.
Speaker 2That is actually one of the most high demand spaces that I get. So, like tomorrow, I will be at a manufacturing space doing a keynote for a large group of businesses. Space doing a keynote for a large group of businesses also in the manufacturing, agriculture, hospitality and medical industries. But manufacturing always seems to be huge, yes.
Speaker 1Yeah, I think the interesting thing about AI at the moment for the last 10 years nearly, we've been talking about the fourth industrial revolution, industry 4.0. It hasn't really delivered on what people expected and it feels like maybe that was because connectivity wasn't there and ingredient technologies like AI were not there to manage that huge amount of data that does come from the factory floor and that can be converted to maybe insert and value add there. Is that something you see and do you think? Where we are with AI now is far enough forward that that's the game changer that can push us to that kind of model of self-healing factories much more flexible, high mix, low volume manufacturing and manufacturing that can be competitive in perhaps higher labor rate regions of the world.
Speaker 2Absolutely.
Speaker 2I definitely think that contextual data understanding is this huge part of generative AI when it comes to the manufacturing space.
Speaker 2So in manufacturing spaces, from my experience, data can live across the business in these various silos, and if you go into these spaces and you have data living in proprietary systems, you have data living in data lakes or data warehouses and you have people that are trying to understand how they can now take this mass historical amounts of data they've been sitting on for years and make it into something productive, effective and, as you said, something that can actually change the game for these sectors. So, using generative AI to pull this data together into insights, into predictive forecasting models or used for predictive maintenance I mean the possibilities are truly endless, but it all does come back to data and utilizing data for whatever it is that these manufacturing spaces need, whether it be anomaly detection need whether it be anomaly detection or, like I said, making sense of that historical data to really drive forward where they can now pivot or take that next step into being a more proactive instead of reactive environment.
Speaker 1Yeah, I think the predictability is really interesting and I'm fascinated. We're in a phase at the moment, particularly in electronic manufacturing, where we have a bit of a downturn which is the result of an inventory bubble working its way through post-pandemic. So this kind of what the industry or what supply chain refers to as a bullwhip effect where it's worked its way through the supply chain. There was too much inventory in the EMS, the outsourcing companies, now that's with the OEMs and consumers have slowed down their buying habits. I'm curious to see, kind of next time we see this cycle, if people are using generative AI or AI generally, whether we're going to have greater prediction in terms of demand and whether we're going to be able to connect that kind of those demand signals all the way from consumer right through the manufacturing supply chain, perhaps a little bit like we're seeing start to happen in the in the fast fashion industry, for example. Do you think that's something that AI can help with?
Speaker 2A hundred percent. Yeah, I think that it's interesting. You bring up the fast faction cycle.
Speaker 2A lot of what I do on Generative AI 101 is I try to make sense of how the field of artificial intelligence and then the subfields right of narrow AI and generative AI they play such a role across industries and they can do similar things in similar spaces. So when you start to look at how generative AI can impact spaces like fast fashion and then how you can do this kind of linear frame of how that would impact manufacturing, right from design optimization, so enhanced creativity, but this ability to rapid prototype and then the added ability to just accelerate design cycles, so you've got this now faster time to market. That totally models itself against that fast fashion trend. And then you have the support of material use that can enhance itself or you have a more efficient use of materials. So I always use so when I go into small and medium-sized manufacturing spaces, we definitely talk about that how what's happening in one industry can model into yours.
Speaker 2And when I talk to them about this idea of material use efficiency, we talk about the idea of constructing this analogy of a grocery list, and I will sometimes actually pull up chat GPT, because it's just like the Kleenex of generative AI. Everyone knows what it is. And so we pull that up and I say all right, let's make a grocery list and let's look at how we can prompt it to say I want to make a grocery list for a family of four for five days and I want you to make the list of ingredients, really work together so that there's limited waste.
Speaker 2And I say now take such a simple concept and model it into your space using this idea of the materials you have, the tier one machinery you have, and how you now can use this to creatively prototype to the constraints and limitations of your environment, and I've worked in spaces where we've taken things like 2D image generations, 2d flat. We then move it into 3D modeling software. We inject it with a few different things Sometimes you're working interchangeably between Photoshop and other AI software we inject it with a few different things, sometimes you're working interchangeably between Photoshop and other AI software and ultimately you're getting to a space where you can create this 3D model prototype that can move into a CAD space and become either a 3D printed prototype or you can move it into the assembly, and it's taking a week instead of taking months, and it's taking a week instead of taking months.
Speaker 2So it's all there. But going back to it, yeah, you can really see that correlation between what happens in one industry with AI and how that models into another industry, this being fast fashion into the manufacturing sector.
Speaker 1Yeah, yeah, it's fascinating, isn't it? And you know, as you say, there are those parallels across. Yeah, yeah, it's fascinating, isn't it? And you know, as you say, there are those parallels across. And I think, when I look at the people that are capable of providing support in this, I look at people that have kind of two sets of skills. One is the, is the understanding and ability to use AI and to have that vision of what it can actually do, but the other is the domain knowledge to understand what's actually going on in the factory that we're looking at. Is that the key combination that you think are the skill sets people need to actually make use of this?
Speaker 2I would say so. So that feeds into two things that I often talk about, both in industry spaces and in higher education spaces. So thing one whether you're a faculty member or you're in the C-suite of your organization, you need to understand what generative AI is and what it can do, and if you don't understand that, you're not in the game at all. Okay, the same goes for higher education and faculty. If you don't understand what it can do, you're limited by either stunting students' use of it in your classes or enabling students to use it in more effective ways. But then you get into foundational knowledge. This works in both spaces. You have to understand how your manufacturing space works inside and out. If you're consulting in a space, your manufacturing space works inside and out. If you're consulting in a space, typically when I do work like that, I'm bringing with me engineers or people who are more skilled in some of those spaces, who understand maybe the machine learning components of a tier one machine that I wouldn't understand right.
Speaker 2But you have foundational knowledge, and I think that's one of the things that people have to keep in mind as well, because when you go into a lot of manufacturing spaces as well because when you go into a lot of manufacturing spaces there's that comes with a lot of NDAs and a lot of proprietary knowledge, and so you also have these spaces where people don't always understand that. There are environments where you're not going to introduce generative AI in its truest form in your chat GPTs, in your anthropics cloud, because you're working with proprietary data and proprietary systems. So that goes all the way back to AI literacy. What is Gen AI? What can it do, how does it integrate into systems effectively, and then what are the limitations and risks? So that's all there, and understanding and being able to unpack all that and how you then integrate and inject that into a specific industry is a few different skills, and a lot of times it's more than just one person. It's a collaborative effort across your team.
Speaker 1Yeah, absolutely, and I wanted to switch to that human aspect of it. When we talk about AI, we often talk about this idea of a co-pilot. You know, the skill in using generative AI particularly is what you prompt and what you do with the result and how you decide how accurate or correct that result if it needs correcting. When you're, you know you're working with a university, when you're looking at developing those, those skills, how do you, how do you think about that relationship between ai and the and the person prompting it?
Speaker 2Well, I mean, there's a lot to unpack with that, but so the first thing is with AI. I think one thing that people don't understand about generative AI is there always has to be this human in the loop.
Speaker 2Ok, there's always someone that's actively engaged with or involved with the AI. You're never going to turn on a generative AI system and just let it loose and never come back to it. That's just not how that works. You're always working fairly in tandem with it. So in a university system or in a higher education system, when you're introducing it to the space, you're thinking about general literacy from the standpoint of what are the most basic skills someone needs to engage with AI systems based on what they are, based on what generative AI can do right From text to image, to speech, to video, to music, whatever it might be.
Speaker 2Now we've got what coding is really taken off. 3d modeling is really taking off with the physics that are going into these systems. So now we understand the basic literacy people need to have and then to build on that. That's when you start to look at a framework that's going to bring more industry-specific skills into the fray. So you're going to look at if someone's going into the manufacturing or engineering sector they're going to need to understand maybe more code space and how that works. The 3D modeling space, image recognition software is super beneficial and then you're understanding how that's going to pipe into other AI systems, whether it be machine learning, deep learning, and then the two big ones that are really coming in, that are kind of all part of this mix, are virtual reality and augmented reality.
Speaker 2And those are playing bigger spaces every day in the manufacturing sector and sectors across the industry. So as you can see, as I walk you through it. It's a tiered approach. It starts with foundational knowledge. It's then slowly building and injecting in what are these primary skills? And then what are the skills that are dependent on the industry or focus?
Speaker 1Yeah, and it's having that ability to adapt quickly because things are moving so fast in the AI space. What's been interesting in the last couple of years in the manufacturing space, particularly in the US, but also in Europe, where labor's high cost is the skill shortage and people thinking about using AI to actually grow their business without necessarily growing their headcount as one part of it. But as the markets kind of shifted a little bit, that concern has waned and people have started to think more about what does AI mean in terms of impacting our labor force, impacting the people? What is the effect with respect to one, jobs but two, actually taking the skills that are perhaps tribal knowledge, getting them into the AI and what impact that has? Is that something you think about a lot as you go through these different paths, and is that something people are constantly asking you about?
AI Impact on Job Roles
Speaker 2Yes, no, it is. Anytime I go into a space and I'm every day I'm somewhere talking about something, and now it is really transitioned into this idea of it's always been. Is AI going to take my job? Always, and I always respond yes, you specifically, AI is going to take your job. It called me last night. It told me it's going to take your job a hundred percent. So, but no, everybody have to have this humor to it, Cause you're just you know, it's coming.
Speaker 2You know it's coming as soon as it's, as soon as you open it up to questions. But I think that I try to look at it in two ways. Yes, we are going to see an influx of agentic AI come into spaces. We're seeing trends come out of Microsoft and Salesforce, where there's this ability to integrate and do a lot more with AI systems when you put them together, and a lot of it reminds me of, like we talked about earlier, that modeling of industry and you look in the manufacturing center and you look at robotic process automation and how you can feed all this extensive data into a system and it can learn and it can react.
Speaker 2The same thing is going to happen with AI agents and enterprise systems, where I don't always know if they'll get to a point anytime soon where they'll be smart enough to take jobs away. We've seen some of this in the customer service sector and we've all got really good at yelling like representative into our phone when things get annoying. But will they get good enough to actually start taking higher level jobs? I don't predict any time soon. The one thing I do tell people is, in the United States especially, we love to say I wear many hats.
Speaker 2I wear many hats. It's this point of pride. Oh, I work at this company, but I wear many hats. I wear many hats. It's this point of pride. Oh, I work at this company, but I wear many hats. So I dabble in the space. My hope is that AI and generative AI can get us to a space where you don't wear many hats, where you wear the hat you were supposed to wear when you were hired into this company and the AI takes on the work that you didn't need to take on. You didn't need to take on the marketing. You didn't need to take on the social media. You didn't need to take on the technical writing on top of all the other things you're doing just to be an engineer in the space. Right, you can just have that one hat.
Speaker 2So I'm optimistic that and you have to stay optimistic in some of this, because if you dabble too far into the dark side, it does get kind of concerning. Because if you dabble too far into the dark side, it does get kind of concerning. But as we keep watching the evolution of these systems and we see agentic AI come in, I think we're going to see some wins, we're going to see some shifts, but ultimately they have to be good enough to outpace a human and if we still continue to see hallucinations at the scale that we do, or inaccuracies pop up safety and risk bringing those onto a system, they'll nip that because it's not worth it. It's not worth the one time that your robot malfunctions and causes a situation that's irreparable. So it's give and take, but it's going to keep evolving and it's going very fast.
Speaker 1Yeah, it absolutely is. And when I think about that and I don't want to go too long, but when I think about that what I think is interesting is, as we take that tribal knowledge and we train the AI and we use the AI to do tasks and maybe co-pilot for the next generation as they come through, there's a concern that we perhaps don't train some of those new people in the tribal knowledge because we're training the AI and we're relying on the AI. So I think we're yet to kind of figure out exactly how we work this triangle of the gray hairs with the expertise and the tribal knowledge alongside the AI co-pilot, alongside the new recruits that are coming into the industry, that need to go through some of those processes for training. So I think there's a fascinating mix there going forward and that's part of the development and educational process.
Speaker 2Well, a solution for that is AI customized training. So a lot of spaces are actually engaging with the onboarding of generative AI solutions to provide that customized training, personalized knowledge sharing, so you can actually take with systems like HeyGen Synthesia so these are text-to-video generation systems you could provide someone customized training, you could create documentation that walks people through things, and then the benefit to that is because the whole GPT a generative, pre-trained transformer a transformer was originally developed by Google to be a translation model.
Speaker 2And all the models we have now are just variations of that. So we have this tech that's prime for multilingual environments. We can not only train, we can train in any language we need to. We have the idea to, or the ability to, craft images, but also use generous AI through augmented reality, through things that you know. The next thing we're seeing is vision assistance. So we have people in manufacturing or picking spaces that are wearing. A team viewer has one that I've seen. That's really great. You wear it. It's an augmented reality system and it can actually help you identify different parts on a machine, different products you need to pick. We also now have this space where any company can take that tribal knowledge and integrate it into a RAG system or a retrieval augmented generation system, and you have your own in-house proprietary knowledge base that anyone in your company could naturally language, engage with. And, as we see, these systems continue to evolve. You don't have to just sit and type into it, because I think typing is totally archaic.
Speaker 2Eventually, we'll get to a point where all these systems will be multimodal and have a voice mode and your company will have a voice, brand and a persona that anyone could sit with and chat with and train with at any given time. So manufacturers can use AI systems to just rapidly showcase a new hire how to service a machine, or they could use augmented reality or virtual reality, like I said, to help really show them how they're going to go about fixing something, how they're going to go about understanding operational procedures, so on and so forth. So these tools could also circumvent the training issues.
Speaker 1Yeah, and it becomes that customized tutor I'm constantly being bombarded with makes you fluent or jump speak or any of those companies that have customized language tuition. But the idea of customized tuition I think is, you know, is hugely valuable that they can teach someone at their own pace. They can teach someone in their own style. People have different communication styles, people learn in different ways, so it's it's hugely valuable. Emily, I'm going to take a leaf out of your book and not go on for too long. Thanks so much for spending time with me today. Absolutely, keep doing what you're doing. I've got I think I've got three of your episodes on regulation and so forth to catch up on. I've been bogged down listening to pre-election podcasts for the last week or so, so I'm going to take some time to catch up on those. But in the meantime, thanks for talking to me and look forward to chatting again soon. Thank you.
Speaker 2Thank you, it's been a pleasure.