AI in Action: Transforming Mining Industry

[Music] Welcome all. Um, Canada is one of the hubs for AI technology and a well-known center for mining industries. So, this event is meant to bring these two industries together and explore how AI can benefit the mining industry. Our panelists are leaders in uh mining industry. They are um promoters of the innovation and adoption of AI technology. So, welcome to our panel. I'll ask panelists to introduce uh themselves, starting with Ingrid.

Ingrid: Hi, I'm Ingrid Hibert. I'm the president of Palango Exploration and on a number of corporate boards. Been in the mining industry my entire life because my father and mother were in the business before me. So, I was coloring maps at six. So, uh, maybe AI can put a do away with some of that. Uh, and that's, uh, that's me. Thank you.

Ingrid: Daniel, I think you're muted.

Daniel: Hi, I'm Daniel Gin, founder and CEO of Six. We're most well known for our flagi platform that we're using here, where we've already embedded a lot of AI features directly into it, using things like automating descriptions, coming up with titles, creating images. Less well-known, but equally interesting on the AI space, is that we built a number of additional AI applications. We've created an app called Photo, where you take a photo of your food and it'll automatically analyze everything that you're eating, determining the nutritional information of it to make it much easier for you to keep on top of your health. We built a platform called ByZero, where you can interact with it like ChatGPT, but it is actually embedded with all of the different AI providers. So, instead of being limited to one provider like ChatGPT or Anthropic, you can actually go and rotate between all of them and use the latest models. And so, there's a lot of work that we've been doing in the AI space. There's a handful of other companies I might mention as well that we've been pioneers on the AI innovation of. Thank you for putting together this excellent.

Ingrid: Thank you, Daniel. Very exciting. James, uh, my name is James Carich Chism. I'm a president of Eureka Maps Incorporated. It's a company that has created worldwide maps for things like gold, silver, platinum, um, molybdenum, uh, basically for targeted drilling of of for mining, for accelerated, uh, exploration.

Ingrid: And, uh, thank you, James. And Kyle.

Kyle: Good afternoon, everyone. My name is Kyle Mcll. I'm the director of commercialization services and member relations with the Mining Innovation Commercialization Accelerator. Uh, we run a national network that's here to support the development of new technologies and we support their, uh, basically commercialization and adoption by the mining industry. Uh, so, very excited to be here and have a conversation about AI, uh, and its, uh, applications and the path to adoption in the mining industry.

Ingrid: Excellent. Thank you, Kyle. Um, so, I'll, I'm moderator Nina Haras. I'm the owner at Theo, a small consulting company focused on the, um, project execution, capital project development and execution, but also on AI adoption, the process of the adoption, the identification of the use cases, um, and so on. So, just for everyone, uh, watching this, uh, uh, presentation, please put your comments or questions in the chat. It's in the right, on the right-hand side of your screen at the bottom. Um, we will start the discussion with a question that had been kicked around, especially the last 10 years. We have seen an increase in duration of the exploration and permitting phase. So, Ingrid, as the owner of an exploration company, how do you see this impacts the investment in mining? And what are the other issues impacting, uh, your progress, your ability to raise capital?

Ingrid: Well, raising capital, that's probably the number one issue facing, uh, exploration companies. And we all know, no, we don't all know, we in the industry have to make sure that the broader world knows. No exploration equals no mines, equals no AI, no computers, no nothing. So, the world needs exploration, and we need new investors, younger investors coming up. And we need to shorten the time because one of the issues for investors is the timeline is so long. We're in a stage now where a lot of the easy deposits have been found. So, we're looking, you know, we like to say, sort of the second layer of the layer cake, a little bit deeper. So, we need, uh, new technologies to help speed up, uh, you know, the exploration stage, and we also need technologies that can reduce costs. And I think AI has can play, I'm not sure we're there yet entirely, um, but I think AI can play a big part in that. Uh, our first try with AI was about five years ago, 2019. And I'm anxious, we're about to redo it, so I'm anxious to see how things have changed. We used a consulting company out of Germany, and it was interesting, and it was good, but it wasn't game-changing yet. So, but when you see how quickly things are changing right now, I'm, I'm really excited for the future. And the other thing I think it can do, and we were talking about this as we were waiting for this to start, is if you can attract some of the people who are interested in AI and they can, and new technologies, and they can make the connection between that and mining, the mining industry, it may be a route to a whole new group of investors.

Ingrid: Very good. Thank you, Ingrid. Uh, James, I'm sure you have something to add to this.

James: Well, Ingrid is right. Uh, like for instance, in, uh, Canada, to get an average mine going is, uh, after exploration stage, even, is five to 10 years. So, investment just makes a big circle around Canada with those sorts of numbers because, and I shared this investment sheet with you that I did on, on, uh, the loss in terms of permit times. By year eight, the average mine is not economic for an investor. So, if you have eight years of, uh, of permitting, which happens in the US and often in Canada, it's not worthwhile for your average investor to make an investment. We have to get to the point where our permit cycle is down to 18 months. And how I accelerate that with data mining is basically to, to approach, uh, exploration as more of a surgical operation. In terms of, I provide maps that, uh, for instance, you can plan roads towards a very specific surgical area, as opposed to a broadcast area where you have to permit several different areas and check out anomalies that are often empty of metals. Using traditional geophysics with data mining, the premise is that you can just go to a specific area, drill that out, start to establish your, your infill pattern, your outfill pattern of drilling, and progress it that much quicker. So, I always say it's one road, one resource approach when you use AI, as opposed to the older fashioned, uh, exploration technique of drilling everywhere. So, if you look at road permits, look at Canada, it takes three to 12 months to get a road permit in Canada, depending on what province you're in. For, uh, to drill out an area, it just doesn't make sense, right? But, uh, if you're spending your efforts in one specific area, you already know maybe what species at risk are in the area because you can forward plan with this sort of AI. You can compare the AI classifications to maybe species at risk. Now, your geologist has a forecast tool to say, well, if I go into that area, I have to anticipate this. But if I go over here, which also has a resource, I'm avoiding potential problems in the future.

Ingrid: Yeah, very interesting the point that you made. Um, in the past, and I heard you saying that, is that actually with the application of AI in the exploration phase, the duration can be shrunk down to three years rather than three years?

James: Yeah, easily. And I've, you know, in Mines and Money, when I, I used to attend those online conferences often, I would find, you know, everybody throws a map on the wall. And if I hadn't enough information, location, shape of of the hills around the area, I could locate with my own maps whether or not their drilling season was going to be good or not, you know, ahead of time. You know, within 10 minutes, I could just shift my maps over and say, I, I don't know if they're going to do so well, or yeah, that looks pretty good to me, what they're, they're doing. So, that technology is definitely not only good for the exploration company, but investors as well. So, it's, it's really new ways. And as you saw from our previous presentation, even on fraud, like I gave the example of Brix, and in that, during that presentation I had with you, and I don't know if Ingrid has been shared that presentation, but you literally could see it was empty of any exploration potential where Brix was 20 years ago. And with the AI, you can avoid that type of fraudulent situation, just like banks use AI on credit card applications to avoid fraud in credit card applications. With AI, you can say, well, as an investor, I don't see anything there, or I see something there ahead of time.

Daniel: I think the dropped on the call briefly, but I'll add that I'm incredibly bullish on AI in the metals and mining sector. I think that in particular, what we've seen over the last five or so years, maybe even seven or so years now, has been what is fundamentally a bull market in gold bullion, where we're seeing the gold price increase. The reason, in my view, that we haven't seen a corresponding bull market in the metals and mining sector is that the all-in sustaining cost of extracting gold out of the ground has also increased due to inflation and so many other factors. What we're seeing in 2025 is that for the first time, gold companies that are in production stage are actually printing money. They do, in fact, own a gold mine, and it feels like it. They have very strong free cash flow, and the prospect of what the gold price is going to do in the future looks ever more bullish. As we even now in the United States have discussions about going and auditing Fort Knox, gold is becoming mainstream. There's more and more emphasis. And so the gold price, look, the gold bullion price looks like it'll continue to do very well. I think in parallel to the improvement on the gold, on the gold bullion price, what we're going to see is a decrease in all-in sustaining cost. I think largely that's going to be through AI, through automation, through the ability of having the ability to extract resources out of the ground in more of a commercially economical manner. And what that means is that the production stage companies will do phenomenally well. And I think subsequently, all the companies downstream of that, on the development stage, and ultimately at the exploration stage, will become a lot more attractive as acquisition targets. And there, to your point about capital raising, I think as these trends move in tandem, it'll mean that those earlier stage companies could raise a lot more capital.

Ingrid: Right. I'm sorry for the interruption on my side. Ingrid brought, uh, an issue of the, and you, Daniel, talked a little bit about that, attracting the tech investors to the mining industry. And with the understanding by doing that, that you actually have a customer for your product, whereas if you're building something general, you're out on the market without hoping the customer is going to understand your product. So, Kyle, you are coming from the commercial side and mention, can you share with us how a good approach looks like to bringing, bringing product to the market?

Kyle: Yeah, there's, we've seen a number of startups kind of be successful in bringing their new technology or solution to, uh, mine operators. But I do want to pick up on a point that, uh, Daniel just made, and that's with the gold miners. So, obviously, gold, at the price it is, is obviously very attractive. And it's been interesting at a couple of conferences, actually, in fact, the one I met Nina at, one of the gentlemen speaking talked about gold kind of being a first mover or an early adopter of a lot of these technologies based on the fact that they've got a little bit more runway, if you would, to to throw around or to kind of experiment with some of these things. And I think that's very important that the mining industry, um, be a little bit more open to experimentation when it comes to these new technologies and these solutions, which is a little bit difficult for, you know, mine operators in an industry that traditionally is a little bit more risk-averse when it comes to adopting new technologies. So, which I think is a bit of a segue into, you know, what do you need to do to attract customers for some of these new solutions? The thing that I would tell startups or the folks looking to bring a new AI technology to the mining industry is you have to create it like any other technology. The, the fact that it's AI is not the value. You know, it's, it's about the value that the artificial intelligence creates for the mine operator. So, there really still needs to be that very strong value proposition there, uh, for the operator to even take a longer look at adopting the technology. The other piece would be making it relatively easy to adopt and integrate and deploy on site. Now, mine operators all have different, uh, rules and regulations, internal and external, when it comes to bringing, you know, new software or new solutions online. But the easier you can make it as a startup or, or, you know, one of these solution providers, the more easily you'll find it get adopted, uh, by the operator. So, I think that's really important. And I just, I, I want to also mention, Canadian companies are very good at doing the research, building the technology, engineering the solution. We're, we're fantastic at it. There's a reason why, you know, Toronto, Montreal are the hotbeds for AI, and we've got a ton of AI talent coming out of those places. But what we don't, and we're notoriously bad at this, is actually the development part. We don't do spend a lot of money on the business development or on the marketing piece. And I think, uh, for any mining technology, regardless of whether or not you've got AI embedded into it, you need to spend the time marketing and communicating what your value proposition is to the marketplace. So, you probably need to spend a little bit more on your marketing and advertising budget than you might, than you might otherwise think you do. Uh, attending conferences, getting your name out there, leveraging platforms like the one we're on today, are all great opportunities for you. And I think you'd really, uh, basically cast a wider net in, in a lot of respects, and easier to adopt.

Ingrid: So, can I jump in here, Nina?

Kyle: Oh, absolutely.

Ingrid: I'm, uh, was and still am an early investor in a, a, a company that is doing AI for the mining industry. And what they did that I thought was, um, brilliant is they went after the OEMs. So, if you can get your AI embedded, like Microsoft has, right? I'm now using, uh, Microsoft, and what do you know, Copilot shows up, and I can do all this stuff inside my universe. Similarly, if you can get your, um, your whatever AI product as part of the, you know, team up with Caterpillar, team up with, that's a, a really quick way. And then you're not, uh, approaching 75 different mining companies. You're so, I thought that was a brilliant way for them to kind of move forward and partner. And it also provided a source of funding for them.

Kyle: So, another model that we've seen work as well is building consortiums around particular products. And we see this more with kind of the hard tech place. So, there's a hardware and software component to it, and typically it's equipment of some kind. But if you can build a consortium around it, get a couple of mine operators in on the same kind of development pathway, it de-risks it for everyone. And if you're able to provide some savings on the back end, you know, when it comes to deployment or on the capex for an initial buy, it, it does, you know, bring, you know, get the interest in the buy-in from the mining, uh, side of things a little bit faster. We've seen a handful of companies be successful with that model to the point where, you know, a couple of them will be building equipment over in the next couple months, and, uh, have a demo track kind of set up to go, all because they were able to bring three mining operators, a manufacturer, and an engineering firm into that consortium.

Ingrid: Excellent. Any other thoughts on this topic?

James: Well, from my perspective, for geology, it's, it's pretty simple. Once you have the classifications, it's basically a geologist can go on the field with one of my maps on his handheld or on a phone and basically walk, even on an exploration, to exactly the classification. Example I give is, uh, I did a map here in Nova Scotia, and then before a P-deck conference, it literally took me 45 minutes to find gold, right? And that's considered impossible in the mining industry. Usually, a group of people, group of geologists, can spend two or three seasons at a field camp looking for visible gold. I, the same day I cranked out the map, I found high-grade gold in 45 minutes, nine grams per tiny gold, right down the road, you know? But when you go to the P-deck, and there's only two other groups in among 17,000 people with visible gold in hand, and you're wandering around saying, hey, I found gold with a new technology, it's a hard sell, you know? So, more examples like that, I'd like to produce for people. I, I do give freebies to people on occasion, just so that they can go and send their geologists out to locations and see what I, if what I'm saying is true. And that's almost where I'm at right now with the industry because your average geologist will not believe that it takes 45 minutes to find visible gold.

Ingrid: Yeah, that's definitely. I'm looking at Kyle, his eyebrows are going up.

James: Yeah, yeah. And I came from the gold industry. I know how long it used to take me before the technology to find gold. It's just not that easy, you know?

Ingrid: Right. So, yeah. So, just, uh, just in case, um, anyone who is listening to this panel, please put your questions in the chat box. We would love to hear from you as well. I like to actually look at the, the ability of AI technology to come in and transform the current processes. We, we are talking here about using AI technology to enhance or replace current methodologies. But with the integration, the processes, and the, I spent lots of time, um, on the automation side of the processing plants. And I believe with the, uh, enhanced automation, especially because we are bringing now in a big way, sound and the vision as a part of the control systems, we can actually, in my mind at least, transform the process. We don't need the same processing configurations anymore. We can improve with the bottlenecks. Bottlenecks in processing, for example, exist because there was an issue with the effluent part, whereas if that issue is resolved, then we can transform the, the back end of the process. So, what I'm trying to say here is we shouldn't be looking only at the application where we are replacing existing technology with AI. We should be looking how it actually transforms the processes and improves the, improves the, um, simplifies the process. Hopefully, we, on the other hand, tend to complicate stuff with new technology. And so, just awareness that it can go either way.

Ingrid: So, Ingrid, you little bit touched upon the, uh, business models. The way I look at this, um, AI adoption of AI technology is actually an opportunity to change. Well, it's not opportunity, but we will be forced, and Daniel spoke about that, forced into rethinking business models. What's, what is driving, uh, certain decisions? How we make money, what is our core business? Uh, there are examples already on the market where an exploration company became the, the tech company, company, and vice versa. So, any thoughts on, on the business model changes as a result of introduction of AI? Daniel, do you want to start here? And we talked a little bit about bringing investments from tech to mining, which will be also a change in the business model.

Daniel: Certainly. I think that if we look at the mining life cycle right now, exploration is a huge part of that phase. And subsequently, one of the biggest things that separates mining exploration from oil and gas exploration is really the much higher failure rate that the junior mining industry has. In the scenario where we're able to use AI, as James was describing earlier, and have a great degree of confidence where the assets are, what's underneath the ground, its probability of success, I think that dramatically transforms the entire junior mining industry from something that is very exploration-heavy into something that instead becomes asset-heavy. Instead of it being who is searching for the properties, it becomes who has the properties, and which of those can you actually develop into a producing mine. And so, I think that's one aspect that AI, if we imagine it at maturity, where it solves the exploration challenge, in one way, destroys the industry, but then transforms the business into being more asset-heavy, which really is just the later stages of the industry we currently have anyways. The other aspect, and I think this is something that Nina, you touched on earlier, is that any exploration company that figures out how to solve a core mining problem, whether it's discovery, whether it's bringing down ASICs, whether it's something else, that knowledge of how to do so is going to be tremendously valuable for itself. And packaging that into a product to sell to other companies is, you know, forgive the pun, a gold mine unto itself. And so, I think that we're likely going to see, uh, some companies that emerge, uh, from that category where they partner with exploration, sorry, with an AI technology, but if they're involved in the development, that AI technology company might be more lucrative to build than the actual project itself.

Ingrid: Right. And we are already seeing that some of the, um, AI tools being used for the exploration actually started to stake the properties, um, because they're too good to sell, so or too good to give it to others. Um, so lots of mining companies do have, as, uh, an innovation arm. So, I absolutely believe that either with M&A with existing, uh, technology company or developing internal product, to your point, Daniel, is going to be another way to enhance the, the, uh, or creating another income stream and leverage ups and downs in the mining industry because we all go, I mean, the mining industry is very susceptible to cycles. So, creating another way to generate income is definitely a way to go.

Kyle: Well, just to pick up on the, uh, exploration side of things here, and, you know, AI moving everything faster, which I think is fantastic. But part of the conversation I think that we're missing here is also how can AI support a mine operator or a junior miner gain license to operate, which is something that we have to deal with a lot here in Canada. So, how can an AI-enabled solution or, you know, equipment or whatever have you make it more environmentally friendly to mine? How can it optimize the life site, the life of the mine? All these kinds of things so that we can gain license to operate much faster, which more often than not is the kind of hang-up to actually getting to production, right? To being able to put, you know, shovels in the ground, as it were. So, I think that's a really important piece of this. We see a lot of technologies that would allow operations to be more environmentally friendly, you know, produce less emissions, you know, optimize their operations in terms of fleet logistics and all those sorts of things. But unless you've got the community buy-in on the front end to be able to put those shovels in the ground, it's, it's still going to take a long time to get a mine up and going. So, you know, part of this should also be how can these AI tools make it easier for a mine operator to gain license to operate? How can it help, you know, with your community consultation? How can it support the dissemination of information around the project to those communities? In terms of business model innovation, how come some of these small-scale mining, you know, technologies also support community buy-in or, you know, the, you know, ownership of some of these new mine operations? So, I think there's, there's there's a lot there too that is important not to miss in this broader conversation about the applications and adoption of AI.

Ingrid: Yeah, absolutely. The, uh, quite a few, um, AI technology companies are actually looking at the environmental side of it. I know James, um, you actually mentioned a piece about roads and the footprint that your particular tool, in addition to exploration, can cover. Do you want to expand on that?

James: Yeah, well, okay. So, for instance, within Canada, it's, it's 30 to 50 square meters resolution on the surface. So, basically, you can park a drilling truck or a drilling rig in that that square and drill it out basically at the center. And so, specific, it's very specific, if you will, in terms of drilling. That's that's not an issue. Another thing that I, I can bring from an experience even in environmental, I did work on environmental in a road job right in the end of the 401. So, was outside the mining industry. We actively held open houses with the communities about the changes we were making. But one of the instances that came up with Wood, actually, everybody here is probably familiar with Wood as an engineering outfit here in Canada, one thing, and relates to mining actually, we had to do a study of about 200,000 plants that were on the footprint of this highway that we were working on. And, um, the biologist said, well, this is going to take about four years to do this, to count all these plants, tag them with GIS. And I said, no. So, I wandered off and in about two hours, I took, when they do these jobs, they have low aerial photos that they take electronically over the roads, right? Same as mining. And I was able to just train an algorithm and do the count. And within about two hours, so it was all surveyed out that quickly. And then I brought back to Wood, I said, and I was a manager on that job, it was like, okay, you, you're saying that, well, I just wandered off, did that in two hours, came back, there's your survey. And it, and the field biologists have a 50% success rate of successfully identifying plants and animals. Nobody ever talks about that. I had a 98% success rate with the algorithm I applied.

Ingrid: Yeah, that's amazing. I'm just looking at Kyle's eyebrows.

James: Well, my eyebrows give me away. That was the biggest thing with these video conferences. But, you know, I think it goes back to that awareness piece, right? So, James, you have that particular skill set, you know how to use the tool, you know to put those solutions together. But that other group of folks that you're working with certainly didn't know that that was a possibility or, or, you know, or even something that they could they could use on on a site. So, I think there's this, you know, that underlying education piece to what can you actually do with these solutions, more than just, you know, ask it to write you an email or, you know, generate a, you know, the world's best cookie recipe or something for me. You know, how can we actually use it in these real-world industrial applications? I think we're still waking up to what's possible with these, but that's that's a tremendous example of I think how how these things could be applied, particularly along the environmental side for a license to operate.

James: Yeah, and then with city jobs, and this was the in Windsor, Ontario. You get everything with your open houses. You get crazies to highly tuned up professionals asking questions. You have media immediately in front of you. So, it's all very political, right? At the end of the 401 highway. So, if we could adopt more of what the highways do, even in Ontario, to deal with the public, as as miners, if, if we take some of their examples of how they approach the public, and they do it quite successfully all the time, I think would help us, you know? Because I, when I'd attend these open houses, they were very well done from the media side of things, extremely well done, very good.

Ingrid: I like to call that cross-contamination between industries. You know, when you go into one field, if you're applying for a job, they ask you how many years of experience you have in that particular field. And I don't know that it is always, I'm sorry, an advantage because you can bring things that are done in the other areas and implement them in this field. And I've seen it. I worked through across the industries and I've seen the value of that. To your point, James, Ingrid, we are, I'm learning a lot. And from your perspective as the owner of the exploration company, I'm sure you're hearing some first time. So, to the point of the having that educational actually need for this kind of conversations to have and educate ourselves about what's out there, what's possible. I think that's that's really kind of minimal that or first need that we have to create or first, um, opportunity we have to create within the mining industry and continue talking about this. Any thoughts, Ingrid, on this?

Ingrid: Yeah, so that's exactly what I was thinking is, uh, you know, we need to, uh, disseminate the stories with the proof behind it, so that, uh, people, people who are going to have to write checks can can look and find them. For one, because right now it seems to be scattered all over. And, you know, we all want a sort of a field test. So, if, and I guess maybe Daniel, you're part of this, is a, you know, a platform that gets these stories out with the background, right? So, that it's not just a story, but it's a story with the hard-based evidence behind it. And you start to get a few more of these, and everybody will get a bit more excited, and we can try it because we've got a, you know, I, there's, what did I say? And I, I don't even know, it's 2,900, how many, uh, companies say right now that they're providing, um, AI for geology to help you target your drills? How do I possibly select one? But if we could have a place where I can go and look and see what somebody else has done, and that, I mean, that's exploration all over again. Is what was tried somewhere else? Did it work? Okay, well, I'm going to try it over here. That's how we've done it from the beginning of time. And so, I think maybe your platform could be a huge impetus for that.

Daniel: Precisely. I think that right now, even among comparing AI models, the way that we in the industry know which AI models are the best is we have a standardized series of tests. We run each AI model through. And I think to your point, Ingrid, absolutely, we should expect that to emerge similarly in the application of AI tools, like which of them are best at actually identifying targets? What is their accuracy rate? How does that, as James points out, compare to a human? If a human's hit rate is 50% and yours is 97%, obviously that's going to be significant. And when Kyle's raising his eyebrows, I was thinking, wow, that's a huge implication. But how do I trust that data? Having some sort of trusted, verifiable source to see, how does this compare against industry standards from the human? How does it compare against other AI tools? I think is precisely the way to go. At least for my view, it's the test confusion matrix, which confuses people, by the way, is a QA/QC step in AI that I use all the time. And it's a good comparison set when you're doing that form of classification. It is not applicable, per se, in the ChatGPT sense where you have an unconstrained algorithm, but for a constrained algorithm, it's, it's an excellent tool. The problem is, it is an education loop that has to be taught everyone to understand what the implications of that that test confusion matrix are, because that's that's my bread and butter. If I don't do well on that, I know the model doesn't work. If I do well on it, I know it will work in the field. So, education there is is key, I would say.

Ingrid: Yeah, exactly. Okay. Um, and that's why this panel is really the first step in in that direction. That's the intent of this panel to bring us all from different aspects and initiate the conversation. So, one of the, one of the questions, I guess, that each of us should ask ourselves is, what happens if the mining industry, especially the mining industry in Canada, does not adopt, um, AI technology and innovation in general? AI cannot solve all the, all the issues, not yet. But what happens? I mean, we have seen the exponential growth in adoption in China, and every now and then there is another video showing a level of automation that we don't see here, frankly. So, my concern would be, are we falling behind even further by not adopting AI at a pace? Or what, what is the prospect? How, how will we mine in the future if we continue on the current path?

Daniel: So, I'll jump in there and perhaps with a bleak statement, which is, if Canadian, let's say, production stage companies are having a much higher AC than an international counterpart, that international counterpart is going to acquire them and rebuild everything themselves and benefit from the gains. And so, very much, I think that what we'll see is, if Canada isn't able to build its own capacities domestically, it will end up being owned by a foreign government or foreign or foreign entity. Now, obviously, we're already seeing the Canadian government take steps to stop mergers and acquisitions out of the view of national security. And in those cases, the result will just be a very non-competitive country where we have mostly zombie companies that are not able to actually hit above their weight, that rely very heavily on government subsidies, and that's falling further and further behind. But that's my optimistic view.

Ingrid: So, do we think that Canadian industries are not adopting because that's not my perception? It's the pace of the adoption. Ingrid?

Ingrid: Yeah, lots of, lots of companies are trying out, uh, to your point, testing. But there is the pace of adoption.

In this, is for me, from my point, it's very superficial what I've seen. Some examples of the adoption in China compared to where we are in Canada, it's, it's not happening fast enough. Okay, I'm going to play devil's advocate for a minute though. Uh, you know, we have regulations in this country for a reason and, you know, we don't want to make to go so fast that we make catastrophic errors. So again, I guess maybe it's about information sharing as much as anything because, um, you know, there are, there is a dark side. There is a risk to AI adoption too quickly, I think, if it's not well enough understood. And, uh, um, you know, I, I hesitate to, to, uh, criticize the Canadian industry for being careful about people's lives, safety, uh, all of those things. So, um, I, but, you know, I don't have firsthand knowledge of it, but the little bit I do see, you know, when I was on the board of Kirkland Lake, they were one of the first underground mines to accept, uh, start with electric underground vehicles. So, you know, know Canadian industry does seem to be quite innovative and and, uh, moving forward with it, but with reasonable caution, I think. But I'm, I'm open to being corrected. I, I myself when I started in mining, and I started underground coal as a, as a geologist, and then, uh, we were on the government side, and then, um, we had the gas explosion that occurred in Nova Scotia at the time, it's Springhill, right? So we had to gather statistics of all the mining accidents over the past. They, they had statistics over three centuries here in Nova Scotia. And of course, uh, I come from a mining family too. My, my Uncle Joe Chisum ran Kirkland Lake back in the 60s, so he was the mine manager there. And, uh, and I personally, in the mine industry, I've seen how wrong things can go if you don't do things carefully. I think I brought up the tailings pond the other night with you. We had a cracked tailings pond. I fixed it at the time, but, uh, the mining industry, if you don't approach things carefully on the production side, especially big things can happen that go wrong in a big way.

Um, unconstrained algorithms themselves are dangerous. Uh, sometimes they can create illusions of data, illusions of response that, that don't relate, relate to reality. I'm a constrained algorithm guy. I like constraints within those algorithms. But on the unconstrained side, if you have a, say, a piece of equipment run by an unconstrained algorithm, um, that algorithm can grow and keep developing in situ actually. And that's a concern that you hear like the Elon Musks of the world bring up all the time. That's a true concern. On, on my side of things, it's pretty simple. If, you know, I'm creating classifications, they're not going to take over the world, other than improve mining. You know, that's all they're going to do. But with ChatGPT, I see all the time, it's falsifying responses all the time. It's rather irritating when I see it happen. Yeah, it's non-deterministic algorithms, so you're going to get all kinds of issues. But that's one thing. When it comes to regulation, the EU actually has the most stringent regulation, but even them are now looking to relaxing them because they understand that it's, uh, more of a burden for the, uh, development than the than the benefit. So it's, it's that play between regulation and, and, and freedom to explore. It's, um, Canada doesn't have one. Uh, the US has the, um, some limited, um, regulation. I don't know about the rest of the world. Kyle, Daniel, can you contribute to this?

So increasingly, what we're seeing is probably the US has arguably the best approach so far towards its AI policies. The European Union has arguably the worst. The result of a lot of the EU regulations are AI changes typically weekly. The Europe typically gets, uh, its AI models three months after everybody else does. They have to go through all those regulations and so on a weekly basis. That means they're 12 versions behind seeing what the latest thing is. Um, in terms of how AI companies are being or being applied, Palantir is probably the most, uh, established company in the United States, which is also public, using AI algorithms for military applications. And so largely their thesis is that they'll get AI to a point where it has some of those deterministic qualities, even though it's a non-deterministic model, where the accuracy of what it produces is sufficient that it solves whatever problems that it needs to solve. That has applications in commerce and others as well. And so the view, I think, uh, rightfully held here is unconstrained AI could cause huge damage. The question is, can we build AI, or will AI develop over time to a point where it can be applied in a reliable enough manner where it's not hallucinating, where it's not creating these, uh, error-based, these rate of errors, and if it does create those, those errors, can it be caught before it hits and affects production in a catastrophic way? And to some extent, in order to go develop those models locally in Canada, there has to be some sort of commercial application. In the interim, there has to be some kind of feedback loop, and it can't be purely academic or government funded. It has to be in the field commercially, right?

And Kyle, you actually talked about that, um, in early on. Kyle, can you hear us? Yeah, certainly. I think, you know, to Ang's point, you know, there is a fair bit of adoption of these, you know, innovations, just not just the AI, but new technologies, new solutions happening. I think what we'd like to see, certainly as a, a group, an organization that works with, you know, 100 plus SMEs developing new solutions is, um, the industry being more open to, I think, that experimentation and, and adoption faster. Um, I know there, there are some, uh, you know, there are rules and regulations for, for good reason, um, but if you look at where the demand is going for critical minerals and where production is for a lot of these mines, um, the gap is widening significantly. And so if there's going to be any, if there's going to be any way to actually make that transition to meet the demand so that everyone can use battery-powered, you know, everything, we just have to find new ways to mine and, and better ways to mine. And I think a lot of that will come from investment and experimentation with a lot of these new innovations. Uh, I think a call to action for the, the mining industry would actually be, can you invest a little bit more? Get more involved in your supply chain? Can you be a partner in the development of these solutions so that they're not being made, uh, they're being made with industry in mind versus, we think this is what they want, uh, and then trying to knock on a couple doors or couple hundred doors, just trying to get the answer. So if the industry is more involved and more upfront about the challenges that they face or the kind of technologies that they're looking to support, um, that goes a long way in helping shape the innovation so it can be adopted in a meaningful, in a meaningful way. And that's something we do with MICA. There are eight mine operators that work with our, our network, and they, we work with their innovation teams to understand what are their challenges, what innovations are they looking for, what do they have budget for this year in terms of adoption, um, and we help communicate that to our members so that it, it can inform their, their product roadmaps, and ultimately ideally lead to better commercialization outcomes for those new technologies. That's, I think, really good, um, end to this discussion. We need collaboration in the industry. Yeah, it's, it's happening, but it's really a question of the pace, just because technology is, um, advancing so fast, and the industry, uh, just is not used to this pace. None of us is used to this pace of the development. So everything is really speeding up. Our whole society changes, and the whole society are speeding up. So it's a new challenge for all of us. So we'll continue talking about that. Uh, hopefully actions, I, I like actions will follow shortly. Uh, somebody us here, uh, part of this panel, um, will have a few ideas. Um, everybody who's listening to this, we invite you to reach out to us and share your ideas, and let's see where we can improve stuff. It's exciting. It's not, we should approach this with excitement, not fear. So that's all I have for the closing. Anything else, James? Well, I appreciate the, uh, the panel and I appreciate folks listening. Um, I do know at this point, China has 90% accurate models. So the problem we also have is that we're going to come to the stage where we're going to find properties bought out from underneath us because the Chinese have moved ahead. They just acquiring the properties. They can see them. 90% models, you're getting to that range now where you're, you pretty much know what you're going to grab in terms of land packages. So, um, that's, I papers right now shown that they have that accuracy level and they're just moving ahead and doing it. You know, yeah, I right now I have 95 to 99.9, but that that little road and then, you know, because my, my main, main competitor is basically a state government. So, yeah. Ang, anything from you? Yeah, thank you. This, I think I've learned a lot and I, I know where to go for some information now. Wonderful. Kyle, I just want to say thank you for the invite. Nina, uh, Daniel, thank you for using our platform. This has been great. I really enjoyed the conversation and, uh, yeah, there's, there's lots of fun ahead for the mining industry, that's for sure. Is that, uh, thank you. Dani, much. Yeah, thank you everyone for joining us and I mean, thanks so much for hosting us. [Music]