[Speaker 1] Good morning, Steve. Good morning. It's late evening for you, isn't it, Phil? [Speaker 6] Yeah, fairly, but, you know, we get by. I'd like to share something with you, if I may, very briefly before the meeting starts or after the meeting, whatever, that I'm in the process of trying to register DPP as an IANA link relation type. [Speaker 1] I saw your email. Yeah, good. [Speaker 6] And they're pushing back on a few things, and I think I've now got the registration right, and I plan to use UNTP as the primary reference and a note to say that it's equivalent to the GS1 link type of DPP. Right. And they're pushing me for a bit more on the definition, and so what I want to suggest to everybody is a definition for the link relation type is a link from a context, that's their word for the anchor, the actual URI that identifies the product, a link from a context URI that identifies the product to its digital product passport. That's it. [Speaker 1] Yeah, you want to keep it simple, don't you, and precise, really? [Speaker 6] So that's what I'm submitting, unless you or someone else on the call says, ah, no, don't do that. [Speaker 1] No, no, I think it's good to have it registered as an IANA link type. That means it's short and precise and universally understood. It's good. I was going to ask whether we should push our luck and ask for the other credential types, like it might be pushing it, you think? [Speaker 6] I don't think it's pushing it, but I'd like to see how this one goes. [Speaker 1] Yeah, okay. Fair enough, let's do that. [Speaker 6] So my thought was if this one gets accepted, which I think having made the changes I'm making now, it will be, I hope so anyway, then if that goes okay, then we can say, ah, right, well, now we've done that, hey, here's some more things we could potentially do. Okay. [Speaker 1] Would it help for me to write anything or for the UN Secretary to write anything? [Speaker 6] If I need to, but let's leave that. That's the link to the now resubmitted request. If anyone wants to make a comment on that, they are more than welcome to do so. It's an open GitHub repo. Anyone can comment whether you want to make an amendment or support it, whatever you want, there it is, if anyone wants to do that. [Speaker 9] All right, thank you. [Speaker 6] Thank you. I'll stop now. That's a good move. [Speaker 1] All right, well, it's four minutes past now. I think you've just consumed, Phil, very nicely the time that we normally sit and wait for the laggards to join anyway. So perfect. Thank you everyone for joining and apologies you haven't seen me for a few weeks because I don't know why I book flights without thinking at exactly the same time as these meetings were on. And note to self, don't do that again. But thank you to Pazak for standing in. Do the usual preliminaries. This is a UN team meeting. Your contributions are considered UNIP. So if you don't want to contribute something that you consider your IP, then don't do so. This meeting is being recorded. If you've got any objections, please let us know, but we've been recording all of them. And the recording will be published publicly. And this meeting is a status development team meeting, not a commercial meeting. There's no promotion of individual products or anything like that. With that, let's launch into it. We've got a couple of things on the agenda today. One is I wanted to share with you and get your input on some thorny chain of custody problems, using as an example the copper supply chain. That'll probably take at least half an hour. And then we have a little bit of a show and tell on what's called tier three testing, which is not just testing that a particular credential is valid, but testing that a combination of credentials makes sense because their context is related. Michael will lead that demo. And also, of course, the announcement that I made earlier that we will spin off these two working groups. And so we won't be talking much about conformity credentials or adoption anymore in these meetings because they'll be in other working group meetings. And I hope to spin off two more so that this one just becomes a sort of a coordination meeting. All right. Has anybody got any preliminaries they'd like to talk about? Have we got anybody on this call who has never joined it before and would like to do a 30-second introduction of themselves? No? OK. Then I'll find that PowerPoint that I sent around. [Speaker 9] Not everyone would know Nancy, perhaps. [Speaker 1] Well, Nancy's joining for a bit. Well, go ahead, Nancy, while I find my PowerPoint. [Speaker 5] Sure. Yeah, hi, everyone. I'm Nancy Norris. I am the chair of the Bureau for UNC FACT, which is, for those who don't know, the standards organization that is the overarching organization that's overseeing UNTP as a technical standard. Yeah, the Bureau, Steve often describes it as a board for a company. And so we set strategic direction and oversee the projects that are implemented by all of our wonderful experts, volunteers. And then we report out to a plenary, which is made up of member states, so very much like our shareholders. And they give us direction on what they want us to focus on and approve certain deliverables. Yeah, happy to be sitting in today. [Speaker 1] All right, and also especially because Nancy has a special interest in a copper industry pilot because there's a lot of copper mined in British Columbia. And today, we're going to look at the copper value chain upstream and try to get you to help me get my head around how these different kinds of chain of custody work in this particular value chain. So here's a diagram. Can everyone see this diagram? Let me put it on full screen. That'll probably help. There we go. Can everyone see that? Yep. Yep, I'm good. Yeah, OK. Good, OK. So we've been talking for a while about how UNTP traceability works, namely that each thing in the supply chain has a passport, which makes its way with the products to the next step in the supply chain. So this diagram shouldn't hold any great surprises. But just so we understand our language for people that aren't that familiar with the copper value chain, it's mined as rocks. And like most mines, the stuff you dig out of the ground has a lot of impurities and a lot of other bits of rock mixed with it. And so the mine site, usually at the mine site, turns the raw ore into something called concentrate, which is still a pile of rocks. It's just a pile of rocks with more copper in it, usually about 30%. So that's what concentrate means, and it comes out of a mine. And then it goes to a smelter, which is a chemical process, a blast furnace refining process, which usually is not at the mine site, but can be. And it produces something called blister, which are like gigantic bars of mostly copper, something about 98%. And then it goes into electric refining process, where what comes out is cathodes, which are usually something like 99.99% pure copper. And then it goes off down the value. Now you've got basically almost pure copper, which gets turned into wire, pipes, God knows what, all kinds of things, and fabricators. And what I've drawn there is some foil. That's just one of many products that get made from copper. And it's an input to the battery supply chain. So there's a battery manufacturer. And of course, batteries go into electric vehicles and then get recycled. So it's a little bit simplified, but it's not too unrealistic. And hopefully, Nancy won't throw any rocks at me for this picture of the copper supply chain. The challenge we have is that at the right-hand side of this, we've got manufactured things that can have barcodes or QR codes or model numbers, batch numbers, serial numbers printed on them, like that coil or battery or the car. But on the left-hand side, particularly at the mine site, we've got a pile of rocks. And so the question here is, how do you identify them? And with a pile of rocks, nobody's putting barcodes on bits of rock. They're usually identified in terms of sale via consignment. It's a batch that comes out of the mine. And the batch number will be part of the consignment record. But I'm basically selling you a train load or a ship load of concentrate. And the blisters and the cathodes usually but not always have some sort of batch number stamped on them. But it may be not a globally unique batch number. You may not always be able to identify who made the thing. And here's the tricky bit. The smelter will make a batch of blister from a mountain of rocks that they've got at the input side of the factory that is sourced often from multiple mines. So we're mixing. So it could be that you've got a mine in British Columbia and a mine in somewhere else. And they've got different sustainability performance characteristics. And if this wasn't copper and it was cobalt or nickel or some other or tungsten tantalum, one of the conflict minerals, you could have a pile of rocks from a conflict zone and a pile of rocks not from a conflict zone going into the same smelter. So there isn't any sort of direct mapping where you can say this blister came from that mine. All you can say is it perhaps came from any number of mines that supply that smelter. And a little bit similar with refining. They buy batches of blister, possibly from different smelters. Although there's often a tighter coupling between smelter and refiner. But then you get the same at the other end, cathodes. Fabricators might usually buy off a market a bunch of cathodes that may come from different refiners. So it gets hard to trace back where these things came from and what composition, what proportion of that cathode came from a conflict zone or came from a high carbon intensity or a low carbon intensity and so on. This is the chain of custody problem in upstream. Are we good so far? How the supply chain works and what the challenges are? Yep. OK. So then the next thing says, well, by the way, there are up to five different models for chain of custody. And I'm only talking about two. There's quite a nice reference from a voluntary standards organization called IRMA. What does it stand for? Institute for Responsible Mining or something like that? I-R-M-A. And they have published a chain of custody standard that you can look at it and it'll describe five chain of custody models, one of which is you force segregation of audited materials. So you only buy copper concentrate from IRMA audited mines. And in the smelter, any material going into the smelter has a separate pathway for IRMA certified mines so that the blister can be said it's 100% from IRMA certified mines and so on. This is an ideal, but it's just not a practical reality. Nobody does that. And then there's sort of an intermediate one. Then there's one called batch, sorry, mass balance is the one I'm showing here, which is the concentrate can come from anywhere, but you keep track at that smelter of inputs and outputs. So that means that with each batch of concentrate needs to come some sort of stock or guarantee of origin that says, oh, this batch of concentrate, this many tons going into the pile, if you like, if you imagine the big pile of rocks at the input of the smelter, has these characteristics from this mine. So you're building kind of stocks at the smelter every time I order, I get some concentrate in. And then I'm spitting out flows, which are the blister blocks. And somebody needs to audit. And I should have drawn this little quota management thing at actually all three of these, but I've only drawn it on two just for keeping space. And any place where you've got mixing, which is all three of these, you need to track what's coming in and what's going out and making sure that there's a balance. And of course, very often, no single customer downstream has a view of, they don't buy everything from a particular refiner or a smelter, they just buy some of the stuff. So how do you know that the statement that you're getting about the intense, let's say, the carbon footprint of the stuff you're buying is accurate? Some independent needs to keep a quota management system of stuff coming in matches stuff coming out. And that's how mass balance is supposed to work, as far as I understand. So this little diagram here is showing that two steps in the value chain would also apply at the smelter. I just needed space on the diagram for what comes next. Potentially, this can be done. Hang on. Nancy's raised her hand. Go ahead. [Speaker 5] Just a question. From your research, would you say that this mass balance calculation, is it like an accounting? It's like, is it a paper exercise? Yeah, yeah. Physical measuring that happens. [Speaker 1] I think it would work in one of two ways. One, it would be done once, and only once, when an auditor visits a mine site and checks historical paperwork. You know, show me what you've ordered over the last year. Show me what you've shipped over the last year. Oh, yes, you've more or less correctly allocated input to output. Or you could also do it on an ongoing, as well as that, you could do it on an ongoing basis. But that wouldn't be a human. That would be just, if you like, every time something comes in, a little ping goes to the quota manager to say, please add this to my stock. And then every time you ship something out, you ping and say, deprecate my stock. And while you're doing it, give me a guarantee of origin certificate for that particular consignment. And then that thing might get audited once a year, or even once every two years. [Speaker 5] Because I think with the Coppermark, and I'm assuming Irma as well, I know for sure the Coppermark, when they're doing their chain of custody standard, the stamp of approval that they're giving is that these companies have the right systems in place to do chain of custody calculation. [Speaker 1] Yes. Somebody was going to say something. [Speaker 3] Yeah. Sorry, Steve, it's Nick Smith. And sorry, I'm not on video. I'm just making the school lunches. I think on this slide, there's actually sort of two things at play from what I've looked at in Australia and Japan. You have some level of mass balancing and some level of calculation, and I think audit protocol on those calculations inside those facilities. So whether that's the refiner or the miner, and that's not dissimilar from having a DCC, a digital conformity credential, some sort of third party that's saying what this agent is saying is true, is actually true because they have some sort of audited system. So that's not necessarily shown here, but there is, you have to do some level of mass balancing inside the system to be able to even know what a blended carbon intensity number is or a blended number. But I think what this slide is also calling out is that then you have a registry, which is, has oversight from the government. And those registries are very, they're different. That's now another system. You're sending that information into the registry and there's an inherent link to the government because most of these systems have some sort of incentive. If it's a low carbon product, the reason why there's a mass balance is because there's some sort of financial incentive. The reason there's a book and claim is because there's some sort of financial stick or incentive. So there's inherently a link to a government authority. And that is for me, that's the bread and butter of where UNTP data goes. It has to eventually go to one of the registries who's doing that kind of balancing. The unseen thing here is that, which is not an established written down practice, but something that auditors are likely to encourage is that the organization that's issuing say things into a registry saying we're booking and claiming these emissions, whether they're saying that they've had those emissions or they're claiming emissions reductions, they'll have to do some sort of time balancing to match with their actual plant site emissions. So you have these footprint emissions and then you have your site emissions and you want to make sure that they work in tandem. The example I can give here is if say the Permian basin is saying that they're producing emissions and producing a certain amount of low carbon methane through the year and they're issuing certificates on that, the emissions when you sum up all of those products they've issued over 12 months, the direct emissions should match what they were also telling the EPA they've emitted, you know, one, two and three. [Speaker 1] So this is similar to what Australian government is starting to do with their guarantee of origin scheme. And that's why I called it a guarantee of origin for want of a better name. And I accept, I think too, that that little database symbol that could live inside the mine and be just a purely internal accounting process, which is audited once a year by an authority, or it could live outside the mine that gets pinged in a fairly automated way and issues an authority signed guarantee of origin with each consignment or something like this. That's the way Australia is planning to do it. But I didn't want to be too prescriptive here, but, you know, at some point somebody has got to check the balancing, right? That's that quota manager and whether it's done every shipment in an automated way or once a year or once every two years in an audit, I think, I don't know. [Speaker 3] I think the shipment is really important as a concept in bulk commodities because that's where you have some of the strictest controls around triple or quadruple measurement of what's moving through the supply chain. And it's often where the transaction points are. There are other transaction points as well. So you can have blending trading, which I think you allude to in next slides. But yeah, the shipment is an important point, which is probably what will get sent to registries. [Speaker 1] Yeah. Yeah. Okay. So that's a picture of the mass balance process. And what it's doing is giving confidence to that smelter and to the refiner and so on up to the next step in the value chain that the input that the stuff they're selling at the granularity of the thing they're selling correctly reflects the combination of inputs. So if I've got some zero carbon input and some high carbon input, what I'm selling is 50-50. And somebody has ordered it to be true. That's what this is all about. The next little complication in this is that the average shipment of concentrate is traded apparently one or two times. And the average shipment of cathodes is traded multiple times on commodity markets, mostly the obvious one being the London Metal Exchange. In between the refiner and the fabricator or the mine and the smelter. And these traders basically sometimes it's a pure paper movement, right? Like you buy and sell a consignment of concentrate while it's on the boat. And sometimes more of this cathode end, it's actually a warehouse movement. So a trader might buy 100 tons of copper cathodes when they think the market is low, stick it in a warehouse and then sell later. And they're very, very, very, very, I can't say enough berries, protective of where they purchased from. Because if you know where they got it from as a buyer, it might give you a clue at what price they bought it at and what margin they're making. And so they basically, if we try to tell them they've got to pass on the upstream identity, then we have to tell the London Metal Exchange to shut down. And this is just another one of those impracticalities, right? So we need a way where these materials can flow through traders and still retain the important sustainability characteristics while obfuscating the sensitive commercial information. And so some people say this is why mass balance can never work at all in the mineral supply chain. I'm not ready to give up on it just yet. I'd rather see if we can find a way to meet the trader's needs and still achieve the outcome that the smelter and refiner and fabricator and so on have some confidence that the stuff they're getting has the right characteristics. What I've drawn here is basically these credentials or stuff that comes out of the mine when they go to a trader. They can redact the identifying information but still pass on the sustainability performance metrics and so on. Nick, you've got a comment again. [Speaker 3] Yeah, I was trying to think through this one for you. And I saw quite a lot of similarity to how the UNTP DPP data works in the base case because I think we've had fairly simple depictions of supply chains so far except for one of the great slides that you put together, Steve, which was going, I think, a bit deeper into the detail. And when one actually sort of starts to map all the data that changes at each transformation or aggregation point along a supply chain, there's a lot of data that will get left behind or referenced but not actually sort of passed forward naturally in a UNTP supply chain. In the chemical supply chain, you could have 45 transformations going from end to end. So there's a lot of different DPPs that would be issued. So I sort of looked at this and thought, well, this is not that dissimilar from how the UNTP would work base case. And I kind of was thinking about those traders. On one hand, they have the right to ask for certain information on what's coming in. The seller still owns the upstream data selling into a trader so they can control what they're providing. And to me, there's a natural pressure in the supply chain. If people in the downstream want the information, that pressure flows all the way up to the traders. So the traders can redact and reissue a DPP. They can take what they've brought in. They can change and reissue a new DPP. And they can redact if they want. But that's then going to have an impact on the downstream customers. So customer pressure would say, well, I'm not going to go to that trader if I want that information. And I know that that's like a bit utopian perhaps. But over time, I feel like that natural pressure of who wants the information and the buyer pressure can actually control what's communicated and not communicated. [Speaker 1] Yeah, that's what I hope as well. And I hope they only redact and don't issue a new DPP because it's the same. It's not a different product, right? It's just a buy and sell operation. So one question I did have is that at the point this concentrate goes onto the market and a trader buys it and he's making a deal with a smelter, he can – maybe the concentrate is a bad example. Let's take the cathodes. This fabricator is buying some copper via a trader on the London Metal Exchange. And they're doing a deal. And they don't exactly know where it came from, but they care that it's 99.998% pure and it's got these sustainability characteristics. And the trader chooses to obfuscate the supplier information. But when the cathodes actually arrive at the fabricator and that batch ID is stamped on them, if the supplier identity can be – or rather the refiner identity can be inferred from the batch ID, then at that point the fabricator knows where they came from. But these batch IDs, and back to the UNTP model, you could have a DPP issued by the refiner that has all the information. It gets redacted by the trader for the purposes of trade negotiation. And it could go through three or four traders before a thing reaches a fabricator. But when it reaches a fabricator, if you can use that batch ID to resolve back, then you've got the DPP as issued by the refiner, not the one redacted by the trader. Is that A, feasible? And is that B, going to be hugely problematic for the trader? Are they going to start demanding that batch IDs get sort of scraped off or something? [Speaker 7] If I may, it would be interesting to see, I would say, in this chain, how we handle the quality file. Because most of the time, depending on the type of product you will make at the end, you need to ensure a full traceability of everything for quality purpose. And for quality purpose, you need to have individual quality files for each step and knowing where it's coming from. So I do understand the issue of the trader. But if you are making a complex, whether it's an aeroplane or a nuclear power plant, you will need to have this traceability. So at least in some industries, that issue has been addressed because of the need of a heavy and hefty quality file. [Speaker 1] Yes, in niche industries like defense and so on. [Speaker 7] Yes, exactly, defense, energy, aerospace, you would have this need. [Speaker 1] Adriana, you got your hand up? [Speaker 8] Yes. We're making the assumption that the information in this diagram that you've got is generally publicly available, whereas one of the issues that we're grappling with in Surpass is user access rights. So can't some of this information be segmented off so that the trader isn't vulnerable, doesn't feel vulnerable, and that it's a subset? So it becomes part of, it's not general knowledge, but it's... [Speaker 6] Can you move your microphone closer to your mouth? And are you... [Speaker 8] So that user access rights aren't compromised, but that it's segmented so that only certain people can have access to that information and the trader's not vulnerable? [Speaker 1] Yes, so I wasn't making the assumption that it's generally public. We do have that decentralized access control idea in UNTP, which could apply to any of these steps. So any of these objects could be public or could be not public. When it comes to access control, though, if, for example, there's some rich information from the mine, which the smelter has a right to know, but the smelter doesn't deal with the mine, he deals with the trader, then the mine doesn't know the smelter. So when you say grant access to someone you know, when you go through four or five trading... Even this fabricator is probably a better example because there's only usually for a particular commodity a few hundred smelters around the world, but there's thousands or millions of these fabricators. So the refiner will have no way of identifying and knowing whether to trust the fabricator, which is why we use kind of a secret key exchange and things like this to do that. But the answer to your question is, yes, we try to deal with that with a decentralized access control protocol. And I'm trying not to overload this story with... This is more about how the chain of custody works and less about how the access control works. Anyway, to continue the story, the last bit that is interesting is another chain of custody model called book and claim, which works like this. It's got absolutely nothing to do with traceability. It's used in agriculture quite a lot. It's emerging. And in agriculture, it's called supply shed. And what it means is rather than knowing and tracing the material that I'm buying all the way to its source and making sure that I'm buying components with materials that are sourced, let's say, from non-conflict zones or from low carbon through traceability, instead, I work at a whole region level, could be a country, where a incentive scheme exists for mines or farms, for that matter, to improve their sustainability performance. And they're audited by some authority. And they say, for example, yes, you've reduced the carbon emissions of your farm or you've improved this or that on your mine. And as a consequence, the mine can book that improvement and get dollars for it. And an OEM right at the other end who wants to reduce their carbon footprint at a corporate level can claim from the same quota. And so it's very similar, I think, to carbon offsets trading. And this is something that I think we'll see more and more of. I know that there are some large companies getting together to do this. Some of the large electronics OEMs, household names, are getting together to see if they can set up these kind of regional incentive schemes for book and claim so that they can claim the improvements without necessarily having to do any traceability. This scheme exists. And often it's positioned as, well, it works just as well because the outcomes are still achieved. Right at the source, there are verifiable improvements being made, which is what we all want at the end of the day. So why don't we just do book and claim and forget all this complicated mass balance stuff? It is something that's sometimes here. So what I wanted to try to do is position these two, not as competitors, but as just maturity steps and be able to articulate when and why it makes perfect sense to do book and claim and when you might need to do mass balance. And some of that is already written into things like the Irma chain of custody model. There are rules about it. They don't specify you must use this chain of custody model. They just describe the five and say, make sure you tell us which one you use. So one of the challenges here is to have a good argument about what are the cases for traceability and transparency and what are the cases for book and claim and how they complement each other rather than compete with each other. And that last bit, I still don't really have a good solution for. But I think it's important to represent book and claim on the picture because it exists and because it has some value. So anybody got any thoughts about book and claim and how it fits with mass balance? [Speaker 5] I'm just wondering whether book and claim is a way, because from the perspective of mining operators, one of the barriers they find to making improvements and more investing in responsible behavior is that they don't actually see any kind of price uplift or preferred contract as a result. And I think it's a result of having such a complex supply chain between the mine and the OEM, say, so that there's just not that trickle back effect of the OEM wanting to pay for more responsibly produced products. And then each actor backwards along the supply chain is willing to pay a little bit more. At some point in the supply chain, that actually just kind of fizzles and it never seems to get back to the mining operator. So there isn't a commercial incentive for them. There's sort of reputational incentives to become more responsible, but not necessarily financial. So I wonder if book and claim is a way to... As a system like UNTP gets adopted and matures, I think there is more potential for that trickle back effect. But in the meantime, perhaps a book and claim system is a way of providing a little bit more direct benefit to a mine. [Speaker 1] Yeah, look, I think it's a well understood problem that whatever price uplift exists at the OEM end, because somebody wants to pay more for more sustainable goods, it's very hard to push that all the way back to the primary production end. If you replaced all this with farming and cotton T-shirts or leather shoes, you'd have the same problem, right? Which is a consumer might pay more, but that premium rarely makes it way all the way back to the farmer, which is exactly the place where the biggest environmental footprint happens and the greatest change is needed. So it's a kind of a perverse misalignment. And so book and claim can overcome that by going direct to where it matters. So I see a lot of value in it, but here's another question. If let's say both are operating, would it be... And that mine is booking improved performance and then it's selling through the value chain improved performance. Is that double counting or is that just more incentives for sustainable behavior and all good? I'm not sure I know the answer to that question. [Speaker 3] Steve, it really comes down to the incentive programs that are set up by the different governments. You can have double counting on a pure emissions registry basis, but we're never really gonna have one emissions registry for a whole supply chain. So it's really the role of the... What you've got is like the audit or the kind of like the registries in the blue boxes here to establish whether or not there's double counting. I don't think that's necessarily the role of UNTP to solve all of the rules of the different organizations putting a book and claim program together or a mass balance program together. They follow some logic, like there is traceability in book and claim. Booking claims sometimes used because of physical constraints in the supply chain. So they're trying to lower the cost of decarbonizing saying you can blend materials together or you can have some location geo separation just to try and stimulate the market. I don't see really a problem with UNTP. UNTP can handle data transfer in both cases. So the registries and the design of the incentive programs will be what they'll be. And UNTP is flexible enough to carry information for the top program as well as the bottom program. I think the point you're calling out on this slide is that it's maybe just a little bit on the complicated side for the top program for a book and claim program because you could probably just do a standalone calculation that has less rigor from point to point. But, you know, that doesn't say that UNTP can't do the calculation needed for the OEM. [Speaker 1] Yeah, I think you're right to call out that there's boundaries on what UNTP can do. So the next question out of all this is, okay, if that's a reasonably useful reflection of how the value chain actually works and where book and claim and mass balance fits, then what does that mean for UNTP? Because obviously we're not globally auditing this stuff. We're not running these quota registries. We're really just an interface standard, right? So for me, it means what UNTP thing is interchanged with these quota registries and moves along the supply chain so that, you know, that guarantee of origin credential could be just a type of digital conformity credential and maybe it's worth profiling one to say this is how it works. And that stocks and flows sounds to me like traceability events, right? And so if I were to overlay UNTP on this, I would be tempted to say, well, you know, we've got a DPP. We know what that is. This guarantee of origin thing is a special kind of DCC. The stocks and flows can be triggered by traceability events as we've defined them. So actually, you know, it's not a bad fit, right? So I'm just trying to, the purpose of all this is to then go, all right, how does the UNTP bits and pieces map to this so that it's more standardized globally? And that's where I was going with all this. Not necessarily to try to solve world hunger, but put a standards viewpoint on it, right? [Speaker 9] Well, what you just said sounded right to me, Steve. The DCC concept already contains verification as one of the conformity approaches. And I think that's what this is. So I'm okay with that. [Speaker 1] All right. Well, so I just want to leave 15 minutes for Michael to take us through tier three testing. So the question for you all is, is this sort of good enough to put an update on the chain of custody page on UNTP with these diagrams and stories and then say, and as a consequence, this is where UNTP plays a role and maybe put some little icons on this diagram that says, you know, this thing is a DPP, this thing is a DCC, this thing is a DTE. And then we can criticize it from there. Ali, you've got your hand up. [Speaker 4] Steve, I just brought the mass balance and the value chain all the way to the mine. I can say we already have the mass balance daily and not doing mass balance. We have it reconciling the data every day and at mine and in the refinery, smelter. So all those data is available and it's not just for auditing and governance, it's just for the performance measure. And all of those data is accurate enough to rely on that because at the end of the day, all the shareholders look at these results and look at it, how much the value is and how much loss we have in any portion of that value chain. So we don't have any problem with the mass balance and in terms of blending happening sometimes after concentrator to smelter or downstream on, let's say, after refinery, we still can track back to the mine in some extent, but we need to give a little flexibility to play the smelter. So we get the concentrate from different sources and they are going to blend in the stockpiles. We don't have one stockpile with one sort of the concentrate. So the concentrate from different sources come in there and blend. And at the end of the day, you will feed the blend of different mines of product to the smelters. So what we could see is we can classify the mines based on the ESG scores, necklace A, B, C, D. And if you don't want to disclose the name and support somehow the traders, this is a client from the mine class A and that's the ESG scores and whatever related to that. And this is the spec of the materials and we can do it every step downstream. And then for the materials, it's the same. So mine, let's say, class A, they are producing A1, let's say, class. So the materials that goes to the smelter B. And we can, instead of just specifying the name, just give a little freedom to the players to just hide some of the information and disclose the others. And one other thing we need to keep in mind is whatever is producing at the mine doesn't end up to the end user in OEM. So we produce, let's say, if we produce 100,000 tons copper, pure copper, the one that did end up at the OEM, in the best case scenario, we're talking about 90,000 tons. So 10,000 tons, we have waste in between. So we need to address that as well, right? So basically, what I'm trying to say is if we rely on the players more and give them a little more freedom, they can use some of those information to help this chain of custody traceability make sense and fit to their operation. [Speaker 1] Yes, sure. I think the management of waste and things like that would be inside these registries, right? So the stocks and flows would need to accommodate time variance, waste, and things like that. So not so much for us, but these registry things would do that. And as for the class A, class B, is there a difference between... because that imposes something on the issuer of this guarantee of origin to say it's coming from class A as opposed to it's coming from Copper Mountain in British Columbia. Surely, if we give the trader the tooling to hide the name without breaking the integrity of the data, then it achieves the same outcome? I don't know. [Speaker 4] I guess so. So basically, what they needed is, okay, it's coming from the reliable source classified as let's say class A. So it doesn't matter if it comes from the Red Crease operation, Copper Mountain, or from the other one. [Speaker 1] Okay. So a classification scheme could be handy. Maybe International Copper Association would have something to say on that. All right. Where are you from, Ali? [Speaker 4] Canada. [Speaker 1] Okay. Okay. Good. All right. Well, look, let's park this now. There'll be more discussion, and we can do it online and so on. We've only got 10 minutes left. Over to you, Michael. Is he here, actually? Maybe I'm handing over to someone who's not here. That would be funny. No, I don't think he made it. All right. [Speaker 11] No, he is here. [Speaker 1] Oh, is he? [Speaker 11] Yeah. [Speaker 1] Oh. Well, we can't hear him yet. Can't hear you, Michael. [Speaker 2] Trying to fix that. Can you hear me now? [Speaker 1] Yep, there you are. [Speaker 2] Okay. Hello, everyone. I'm Michael. So I'll just spend the last few minutes giving an overview of a proof of concept that we've been doing for Tier 3 testing. I'll move quite quickly. Feel free to ask questions and interrupt. So up until now, we've got Tier 1 and Tier 2 testing. I'll just explain briefly what they are as we're also testing the tool out. So I'm running this command on a command line here, but that was just for the proof of concept. We can just as equally run this in a web browser or in a person's mobile phone web browser, for example. But initially, I'm just going to run it with one credential, which is a product passport, which I've got up here. So this product passport has been issued by a battery company, and it has a few claims just about where are we? Emissions. So we've got claims about battery assembly, one about battery packaging, and one about battery disposal, for example. So if I run that just with that one credential, it will go through and do the Tier 1 testing, which is just ensuring that the credential is valid JSON, valid JSON linked data, and it's a valid verifiable credential. So it matches the verifiable credential schema. The Tier 2 testing, which is where we identify what the credential is. In this case, it's DPP, and we verify it against our own DPP JSON schema, make sure it fits. That part of the testing is already done in the UNTP playground. And just as an aside, we are intending that once we have this proof of concept at a point where we're ready, we want to move this into the playground so it will all be there. The difference is this Tier 3 part. So what happens here is that the credential that's been passed gets added to a graph. And this is where we get the benefit of using JSON linked data, where we can actually just pull these different credentials from different places on the internet and put them in a graph and analyze and actually infer more meaning from that. So in this case, it just comes back and says, well, sorry, I don't trust this web battery company. As a user, I could say, well, I trust them. I know who they are, so that's okay. In this case, we've actually got another credential. So I'm going to add a digital identity anchor for that DPP, which I've got. So here, that's just a pretty simple credential, which in this case, it's issued directly by a government department just to make things quicker. And it says that they know who this battery company is and they trust that decentralized identifier. So I'll run it again there. And this time, again, it has to do the Tier 1 and Tier 2 testing on both of the credentials that have been passed. And it comes through to the Tier 3 testing and says, okay, the only unattested person at the moment is this government organization. Obviously, at that point, we could have it built in that we just trust that. I've just made an option here where I can say, well, yes, I trust that. That's a government one. So again, run it a third time with those two credentials and saying, I trust the government department. And it comes through and says, great. Okay, so we've got a battery. We've filtered out one of the unattested issues, which was the government one, which I've said I trust. So we now have all issues are attested. There's two claims that it's identified and with criteria. Now, none of the criteria have been attested to. So they're just claims that the product has made at this point. I'm just going to skip ahead just to save time. And so I've got a bunch of other credentials and what we can actually do is instead of passing them individually is just pass a directory full of credentials. Or if this was in a web browser, you would, you know, a person would just upload a bundle or even the, the tools that we've chosen, which I may get a chance to chat about briefly actually allow following links. So a credential may have a link to another credential and it can actually download it and so on. So in this case, you'll see it's now got a whole bunch. It's got a conformity credential in which it's been passed an identity anchor for the conformity credential. There's then an identity anchor for that digital identity anchor, as in it wasn't from a government agency. So it was three steps involved there. And what it's got done is it's after doing the T1 and T2 testing, it's created a graph. It's run the inferences that we've got set up, which do things like verifying that the trust chain of issuers goes back to something that is actually trusted, which I may briefly show in a tick. And then it's also compared to the criteria of those claims and match them together. So now you can see that for that first claim that the product makes this first product, one criteria is actually verified by the electric electronic certifier conformity credential, which we have, whereas the other isn't. Whereas the second claim that the product makes, both of those criteria have been verified, attested to by that certificate. And so it can give you a summary at the end. So this would work just the same in a web browser on a mobile phone in that, yeah, you could give it a bundle of credentials and it can do this analysis. Now, the interesting thing is the technologies that were chosen, they don't actually require us to run some infrastructure of, you know, that allows graph querying and inference and so on. It all just runs directly in a web browser through the magic of kind of some new technologies in the last few years. So to give you a quick idea, so the inferences that we're running, this is actually just a, a, a superset of RDF, which in a language called notation three, it was originally something that Tim Berners-Lee came up with and it's been since taken on board by other people. And there's quite a community around it, but it allows us to do these query, these inferences. So I've set up three inferences currently. I'll just briefly overview what they do. So this first one just says, find all the criteria of claims in digital product passports, and then find all the criterias that have been assessed in digital conformity credentials, and then only match them. If the products match the topics match, the conformance has been assessed as true. And so criterians match. And if so, it just adds, it adds a little inference information to the graph saying that, yes, this, this specific criterion has been attested to. We then have a second, here's another inference that I've added explicitly, which basically says for each claim, if all the criterion have been attested to, then make that claim something that also has been verified. So it adds that information that that claim has been verified. And the third one, and probably more important for our purposes is the trust chain. So this one basically says, find all the credentials in that you've been given in the graph and actually find out which credentials have been a test, which credentials have a digital identity anchor and add information saying that that credential has a digital identity anchor. And we use that to then know which ones don't. And then so we can, the application actually actually follows those chains through back to the, the, the government department in this case, which I've said I trust. Yeah. And lastly, so it is set up in a way that anyone, a user or a company could drop their own inference in here as well. And that would automatically run their inference as it runs this process, which might, for example, highlight, highlight claims that don't match certain policies that they have, for example. So it can also be adaptable. There's a clear separation of what the application's doing in terms of creating an IDF graph that can be analyzed and using this separate interface where the inferences can be added. So information can be added to that graph. And then finally being able to extract the, I guess the data that the user wants for a user that is, can I trust this information and does it meet, you know, my environmental kind of standards that I personally have any questions? Nope. Feel free to reach out if you've got any questions or want to want to know any more about the technology, I'd be happy to share it. And yeah, to enable other people to, to join in and help. [Speaker 6] So just, just very quickly, that was fascinating to see I may well reach out and have a separate conversation with you when we're not at the end of the day. Definitely. [Speaker 1] Yeah. That was a lot of technical code going through the screen, but it's extremely valuable when you think about it because it's, it's the key to getting trustworthy data. It's not within one credential. It's the relationship between credentials. So when this goes onto the playground, do you think we're going to be able to see a sort of a pretty little graph version of this where I start dropping credentials in and it draws nodes. And then I don't know, maybe I have the sort of red, amber and green lines between nodes about how verified they are, this sort of thing. Some sort of visualization of. [Speaker 2] Yeah, absolutely. That's one of the next things I'm going to do is, is to actually, I realize it's a CLI proof of concept for, for software engineer, but, and, and I'd like to make it more visual. I don't think I would want to do actual nodes of the nodes in the graph, but maybe abstract out a bit to, for example, is to have a trust graph of issuers. So you can say, oh yeah, I'm trusting this battery company and that goes through to this, you know, conformance company. But it then depends on this company and I don't know them and highlighting that with colors and so on. [Speaker 1] Yes. Yes. You're right. The graph could get so messy. You wouldn't see the forest for the trees. Yeah. Every node in it. Yeah. [Speaker 2] And I should say too, there's, it is a proof of concept. So at the moment it's not doing things like I'm, I'm not verifying signatures on the certificates or anything like that. I'm just starting with the actual data that's usable. Brett, sorry, you had a question. [Speaker 10] Oh yeah. It was just wanting to learn if this is like available on GitHub to take a look at, or should we just reach out to you separately to kind of ask more questions? [Speaker 2] It is. Let me put the link. So I've got it right here. It's just on my personal GitHub at the moment. But yeah, as I say, we need, it will be quite a bit of work for us to to, to move it onto into the UNTP playground, because there's a bunch of assumptions in the UI of the UNTP playground, which don't go together with kind of uploading a bunch of credentials. So there's the link and there's a, I've got a pretty detailed read me there that outlines the more about the technical side of things, the tools that we've chosen. If you're interested. [Speaker 1] Okay. [Speaker 2] Yeah, that sounds great. Thank you. [Speaker 1] All right. Well, thanks, Michael. That was, that was amazing. Actually, it's going to get even better. Brett had his hand up. I know we're towards the end. People should feel free to leave if they've got other commitments, but Brett, did you want to say something? [Speaker 2] I think he just did. [Speaker 9] Oh, I didn't have my hand up. I did drop a note in the chat, Michael, and I wouldn't mind following up with him offline on that one. Okay. All right. Sure. [Speaker 1] Well, good then. Well, any closing comments from anyone before we wind up this meeting? If not, we'll look forward to seeing you. Some of you are different times. And the next one is more targeted on Europe. Although I do see we've got some Europeans here. Thank you for staying up late. We'll see you in approximately two weeks' time and online in the meantime. I hope today was useful. [Speaker 7] Thank you. Bye-bye. [Speaker 2] Thanks, everyone. [Speaker 7] Thank you.