Your Agent Just Authorized What?! — Jay Mok & Ben Coumes, Paypal
Read full transcript 13 segments
-
>> Hello everybody. How are you doing? >> Hello everybody. How are you doing? Um Um Um Does anybody remember the movie Does anybody remember the movie Does anybody remember the movie Terminator? Terminator? Terminator? Uh anyways, it's one of my favorite uh Uh anyways, it's one of my favorite uh Uh anyways, it's one of my favorite uh movies when I was growing up as a kid. movies when I was growing up as a kid. movies when I was growing up as a kid. And it imagines a world where the And it imagines a world where the And it imagines a world where the machines have taken over, right? And uh machines have taken over, right? And uh machines have taken over, right? And uh uh the nightmare scenario here though in uh the nightmare scenario here though in uh the nightmare scenario here though in in 2026 is not that the machines are or in 2026 is not that the machines are or in 2026 is not that the machines are or the agents are launching nukes, but the agents are launching nukes, but the agents are launching nukes, but rather rather rather uh they've uh taken your wallet and uh they've uh taken your wallet and uh they've uh taken your wallet and they've gone on a shopping spree. they've gone on a shopping spree. they've gone on a shopping spree. And they buy like a bunch of crypto and And they buy like a bunch of crypto and And they buy like a bunch of crypto and new bunch of bunch of spanks for you. new bunch of bunch of spanks for you. new bunch of bunch of spanks for you. Um Um Um but the basically today we're talking but the basically today we're talking but the basically today we're talking about how we safeguard against that and about how we safeguard against that and about how we safeguard against that and uh and uh hopefully we can kind of share uh and uh hopefully we can kind of share uh and uh hopefully we can kind of share a a mental model that you can use when a a mental model that you can use when a a mental model that you can use when you're thinking about agent you're thinking about agent you're thinking about agent authorization. authorization. authorization. Uh my name is uh Jay Mock. I'm a product Uh my name is uh Jay Mock. I'm a product Uh my name is uh Jay Mock. I'm a product manager over at PayPal in agentic manager over at PayPal in agentic manager over at PayPal in agentic payments. And payments. And payments. And >> Hi everyone. I'm Ben Cooms. I am a staff >> Hi everyone. I'm Ben Cooms. I am a staff >> Hi everyone. I'm Ben Cooms. I am a staff software engineer on the software engineer on the software engineer on the payment PayPal enterprise payments team.
-
payment PayPal enterprise payments team. payment PayPal enterprise payments team. >> And together we're going to share kind >> And together we're going to share kind >> And together we're going to share kind of like some knowledge with you. Um so of like some knowledge with you. Um so of like some knowledge with you. Um so hopefully hopefully hopefully uh you find it helpful. Okay. uh you find it helpful. Okay. uh you find it helpful. Okay. So uh the the key questions that we kind So uh the the key questions that we kind So uh the the key questions that we kind of like uh start off with is uh in terms of like uh start off with is uh in terms of like uh start off with is uh in terms of like agent authorization is uh did of like agent authorization is uh did of like agent authorization is uh did the human authorize this? the human authorize this? the human authorize this? Um is this allowed right now in this Um is this allowed right now in this Um is this allowed right now in this scope and can we prove it later, right? scope and can we prove it later, right? scope and can we prove it later, right? And we we kind of like try to make it And we we kind of like try to make it And we we kind of like try to make it general, but in our world of payments, general, but in our world of payments, general, but in our world of payments, yeah, did the human authorize this? Uh yeah, did the human authorize this? Uh yeah, did the human authorize this? Uh that could be like a passkey or of of that could be like a passkey or of of that could be like a passkey or of of that nature. Uh is this allowed right that nature. Uh is this allowed right that nature. Uh is this allowed right now in this scope? It's generally going now in this scope? It's generally going now in this scope? It's generally going to be a time bound um to be a time bound um to be a time bound um you know you know token. you know you know token. you know you know token. Um and an amount and possibly could be Um and an amount and possibly could be Um and an amount and possibly could be identified like a merchant or a identified like a merchant or a identified like a merchant or a the actual product intent. Lastly, can the actual product intent. Lastly, can the actual product intent. Lastly, can we prove it later? This is like if we prove it later? This is like if we prove it later? This is like if something goes wrong, right? And in our something goes wrong, right? And in our something goes wrong, right? And in our world of payments it's generally has to world of payments it's generally has to world of payments it's generally has to do with like the disputes and do with like the disputes and do with like the disputes and in that case and how do you can you in that case and how do you can you in that case and how do you can you prove that prove that prove that you know the the human generally you know the the human generally you know the the human generally authorized that transaction. Right? Um authorized that transaction. Right? Um authorized that transaction. Right? Um But we think the way that you actually But we think the way that you actually But we think the way that you actually answer these three questions is really answer these three questions is really answer these three questions is really dependent on the context. You know, dependent on the context. You know, dependent on the context. You know, context is a overused term, but in this context is a overused term, but in this context is a overused term, but in this case what we what we mean is case what we what we mean is case what we what we mean is um um um you know, is it a low stakes or high you know, is it a low stakes or high you know, is it a low stakes or high stakes kind of stakes kind of stakes kind of scenario?
-
scenario? scenario? Um and Um and Um and is this a kind of like open ecosystem or is this a kind of like open ecosystem or is this a kind of like open ecosystem or closed ecosystem? Do the parties like closed ecosystem? Do the parties like closed ecosystem? Do the parties like know each other? You know, people use know each other? You know, people use know each other? You know, people use the term KYA a lot. Know your agent, but the term KYA a lot. Know your agent, but the term KYA a lot. Know your agent, but you know, what we think about in this you know, what we think about in this you know, what we think about in this scenario is is is really about is it scenario is is is really about is it scenario is is is really about is it like an open or closed ecosystem, right? like an open or closed ecosystem, right? like an open or closed ecosystem, right? And in a payments context it could be And in a payments context it could be And in a payments context it could be like, "Hey, you know, ChatGPT or Gemini, like, "Hey, you know, ChatGPT or Gemini, like, "Hey, you know, ChatGPT or Gemini, right?" That's like kind of like a more right?" That's like kind of like a more right?" That's like kind of like a more of like a closed ecosystem because you of like a closed ecosystem because you of like a closed ecosystem because you know, the those agents know the merchant know, the those agents know the merchant know, the those agents know the merchant generally. Um generally. Um generally. Um I like to use a an analogy. I like to use a an analogy. I like to use a an analogy. I like analogies. And the analogy I I I like analogies. And the analogy I I I like analogies. And the analogy I I like to use is kind of like the you like to use is kind of like the you like to use is kind of like the you know, badging into work. You badge into know, badging into work. You badge into know, badging into work. You badge into work in the front desk. You basically work in the front desk. You basically work in the front desk. You basically are are then led into the building or are are then led into the building or are are then led into the building or you know, let's say it's a set of you know, let's say it's a set of you know, let's say it's a set of buildings. You don't need to like badge buildings. You don't need to like badge buildings. You don't need to like badge in every single time to every other or in every single time to every other or in every single time to every other or for every single room because you're for every single room because you're for every single room because you're already within that trusted boundary, already within that trusted boundary, already within that trusted boundary, right? So then when you meet someone in right? So then when you meet someone in right? So then when you meet someone in within that within your your office within that within your your office within that within your your office building, uh, you kind of have some building, uh, you kind of have some building, uh, you kind of have some element of trust, or hopefully you have element of trust, or hopefully you have element of trust, or hopefully you have some element of trust, uh, because some element of trust, uh, because some element of trust, uh, because you're both, uh, employees at the same you're both, uh, employees at the same you're both, uh, employees at the same company that badged in, right? So, um, company that badged in, right? So, um, company that badged in, right? So, um, that's kind of like the analogy I may I that's kind of like the analogy I may I that's kind of like the analogy I may I may use later in the presentation.
-
may use later in the presentation. may use later in the presentation. Okay, so based on those key questions, Okay, so based on those key questions, Okay, so based on those key questions, we kind of think about like, "Hey, we kind of think about like, "Hey, we kind of think about like, "Hey, what's the mental model that we can what's the mental model that we can what's the mental model that we can build off of this, right?" And we have build off of this, right?" And we have build off of this, right?" And we have this like stakes and evidence matrix, this like stakes and evidence matrix, this like stakes and evidence matrix, and we're going to talk about, uh, these and we're going to talk about, uh, these and we're going to talk about, uh, these three different scenarios. three different scenarios. three different scenarios. Um, and so, uh, we're going to first uh, Um, and so, uh, we're going to first uh, Um, and so, uh, we're going to first uh, and you'll see at the top it's kind of and you'll see at the top it's kind of and you'll see at the top it's kind of like the stakes and counterparty part like the stakes and counterparty part like the stakes and counterparty part that I was just talking about the that I was just talking about the that I was just talking about the context, right? Counterparty is like the context, right? Counterparty is like the context, right? Counterparty is like the open or closed ecosystem. And then open or closed ecosystem. And then open or closed ecosystem. And then authority and like evidence is really authority and like evidence is really authority and like evidence is really about how you answer those those those about how you answer those those those about how you answer those those those three questions I had shared in in the three questions I had shared in in the three questions I had shared in in the prior slide, right? Um, so we'll talk a prior slide, right? Um, so we'll talk a prior slide, right? Um, so we'll talk a little bit first about like cloud code little bit first about like cloud code little bit first about like cloud code since that's what most people are very since that's what most people are very since that's what most people are very familiar with, and, uh, basically, you familiar with, and, uh, basically, you familiar with, and, uh, basically, you know, when you as a human you're going, know, when you as a human you're going, know, when you as a human you're going, you know, using your cloud code, uh, you you know, using your cloud code, uh, you you know, using your cloud code, uh, you know, you might be, uh, then setting up, know, you might be, uh, then setting up, know, you might be, uh, then setting up, uh, your your connectors with your uh, your your connectors with your uh, your your connectors with your GitHub or or, um, you know, Jira or GitHub or or, um, you know, Jira or GitHub or or, um, you know, Jira or whatever, uh, linear or whatever uh, whatever, uh, linear or whatever uh, whatever, uh, linear or whatever uh, tool you're using. And, um, you know, tool you're using. And, um, you know, tool you're using. And, um, you know, you're as part of that process you're you're as part of that process you're you're as part of that process you're kind of like authenticating, so that's kind of like authenticating, so that's kind of like authenticating, so that's how you kind of like get that that human how you kind of like get that that human how you kind of like get that that human authorization and consent um, with those authorization and consent um, with those authorization and consent um, with those those applications and to for the for those applications and to for the for those applications and to for the for cloud container act with them.
-
cloud container act with them. cloud container act with them. Um, in terms of the, uh, Um, in terms of the, uh, Um, in terms of the, uh, actual like, uh, scopes, right? The the actual like, uh, scopes, right? The the actual like, uh, scopes, right? The the example here would be that about, you example here would be that about, you example here would be that about, you know, cloud's like, uh, tool know, cloud's like, uh, tool know, cloud's like, uh, tool permissions. Like people are very permissions. Like people are very permissions. Like people are very familiar probably with, uh, the fact familiar probably with, uh, the fact familiar probably with, uh, the fact that you can allow cloud cloud to to use that you can allow cloud cloud to to use that you can allow cloud cloud to to use certain tools, um, uh, deny or ask cloud certain tools, um, uh, deny or ask cloud certain tools, um, uh, deny or ask cloud to ask you, uh, before for before doing to ask you, uh, before for before doing to ask you, uh, before for before doing something, right? And then in terms of, something, right? And then in terms of, something, right? And then in terms of, uh, the action of like a cloud, uh, we uh, the action of like a cloud, uh, we uh, the action of like a cloud, uh, we generally think, um, because it's a kind generally think, um, because it's a kind generally think, um, because it's a kind of closed ecosystem and it's like you're of closed ecosystem and it's like you're of closed ecosystem and it's like you're coding, uh, the stakes are relatively coding, uh, the stakes are relatively coding, uh, the stakes are relatively low here. And so in terms of the low here. And so in terms of the low here. And so in terms of the evidence or proof, you don't really need evidence or proof, you don't really need evidence or proof, you don't really need to have like that cryptographic proof to have like that cryptographic proof to have like that cryptographic proof um, at that point in time. You can kind um, at that point in time. You can kind um, at that point in time. You can kind of just look at like system logs in of just look at like system logs in of just look at like system logs in order to to or you have the ability to order to to or you have the ability to order to to or you have the ability to just revert revert your changes, right? just revert revert your changes, right? just revert revert your changes, right? So, that's kind of like an example of So, that's kind of like an example of So, that's kind of like an example of applying like this mental model applying like this mental model applying like this mental model and using cloud code in terms of that and using cloud code in terms of that and using cloud code in terms of that scenario. scenario. scenario. Okay, so the next Okay, so the next Okay, so the next example we're going to talk about is example we're going to talk about is example we're going to talk about is a more medium stakes a more medium stakes a more medium stakes scenario and why scenario and why scenario and why why we're calling this medium stakes why we're calling this medium stakes why we're calling this medium stakes even though it's within a known or kind even though it's within a known or kind even though it's within a known or kind of closed ecosystem is because it has to of closed ecosystem is because it has to of closed ecosystem is because it has to do with money and and payments. And so do with money and and payments. And so do with money and and payments. And so that's like the shared vault and OAuth that's like the shared vault and OAuth that's like the shared vault and OAuth scope example.
-
So, in this example where the use case So, in this example where the use case is is like hey you're like a let's say is is like hey you're like a let's say is is like hey you're like a let's say a merchant or Trip Advisor, right? And a merchant or Trip Advisor, right? And a merchant or Trip Advisor, right? And you have you have you have a travel travel company and you have a a travel travel company and you have a a travel travel company and you have a lot of great content that you want to lot of great content that you want to lot of great content that you want to monetize. It could be occupancy data, it monetize. It could be occupancy data, it monetize. It could be occupancy data, it could be like reviews, what have you. could be like reviews, what have you. could be like reviews, what have you. And you have a new customer now. You And you have a new customer now. You And you have a new customer now. You have like a trap like travel agents or have like a trap like travel agents or have like a trap like travel agents or like you know agents that that are buyer like you know agents that that are buyer like you know agents that that are buyer agents that are are coming to you and agents that are are coming to you and agents that are are coming to you and you want to be able to monetize you want to be able to monetize you want to be able to monetize monetize your data, right? Through monetize your data, right? Through monetize your data, right? Through machine payments. machine payments. machine payments. So, we work with a partner and Never So, we work with a partner and Never So, we work with a partner and Never mind to be able to enable that that that mind to be able to enable that that that mind to be able to enable that that that use case and leveraging our use case and leveraging our use case and leveraging our they're leveraging our infrastructure. they're leveraging our infrastructure. they're leveraging our infrastructure. Right? So, there is two pieces of Right? So, there is two pieces of Right? So, there is two pieces of infrastructure that they infrastructure that they infrastructure that they that I like to kind of to call out or that I like to kind of to call out or that I like to kind of to call out or primitives that they use that primitives that they use that primitives that they use that as part of the Braintree or PayPal as part of the Braintree or PayPal as part of the Braintree or PayPal enterprise enterprise enterprise infrastructure. One is like the vault, infrastructure. One is like the vault, infrastructure. One is like the vault, right? And the vault by itself, which is right? And the vault by itself, which is right? And the vault by itself, which is storing all these like storing all these like storing all these like payment credentials on behalf of the payment credentials on behalf of the payment credentials on behalf of the on behalf of the buyer agents, on behalf of the buyer agents, on behalf of the buyer agents, on itself doesn't really do much, but in on itself doesn't really do much, but in on itself doesn't really do much, but in order to create order to create order to create what nevermind creates is a a more what nevermind creates is a a more what nevermind creates is a a more eco closed ecosystem, they then off eco closed ecosystem, they then off eco closed ecosystem, they then off you're able to um you're able to um you're able to um offer offer offer uh or offer access to those payment uh or offer access to those payment uh or offer access to those payment credentials through credentials through credentials through OAuth, right? To all those merchants.
-
OAuth, right? To all those merchants. OAuth, right? To all those merchants. So, in our example before we talked So, in our example before we talked So, in our example before we talked about that that travel travel um travel about that that travel travel um travel about that that travel travel um travel company, right? So, by doing so, they're company, right? So, by doing so, they're company, right? So, by doing so, they're able to then create like an ecosystem um able to then create like an ecosystem um able to then create like an ecosystem um of buyer agents and seller agents of buyer agents and seller agents of buyer agents and seller agents uh and have a more trusted uh and have a more trusted uh and have a more trusted environment, right? So, environment, right? So, environment, right? So, um um um in the in the just kind of talking more in the in the just kind of talking more in the in the just kind of talking more about the use case, like the human then about the use case, like the human then about the use case, like the human then is then going to be authorizing is then going to be authorizing is then going to be authorizing authorizing their their payment. authorizing their their payment. authorizing their their payment. Usually, this is a commercial card Usually, this is a commercial card Usually, this is a commercial card commercial use case. So, you're using commercial use case. So, you're using commercial use case. So, you're using like a commercial card, they share it like a commercial card, they share it like a commercial card, they share it with the buyer agent, travel agent. Um with the buyer agent, travel agent. Um with the buyer agent, travel agent. Um then that uh it it also has then that uh it it also has then that uh it it also has scopes associated with that mandate. So, scopes associated with that mandate. So, scopes associated with that mandate. So, that's how you're able to do controlled that's how you're able to do controlled that's how you're able to do controlled authority, but in terms of like the authority, but in terms of like the authority, but in terms of like the actual like dispute handling, we really actual like dispute handling, we really actual like dispute handling, we really uh don't have like a we're not using uh don't have like a we're not using uh don't have like a we're not using like like like cryptographic proof that's being sent as cryptographic proof that's being sent as cryptographic proof that's being sent as part of that that request, right? At the part of that that request, right? At the part of that that request, right? At the end of the day, they can since it's more end of the day, they can since it's more end of the day, they can since it's more of a of a of a closed ecosystem, closed ecosystem, closed ecosystem, they're able to leverage like the just they're able to leverage like the just they're able to leverage like the just the existing the existing the existing transaction logs.
-
transaction logs. transaction logs. Right? So, that's kind of an example of Right? So, that's kind of an example of Right? So, that's kind of an example of like a medium stakes and like a medium stakes and like a medium stakes and use case or scenario, and use case or scenario, and use case or scenario, and we we believe it's medium stakes because we we believe it's medium stakes because we we believe it's medium stakes because of the fact that it is a more closed of the fact that it is a more closed of the fact that it is a more closed ecosystem and doesn't require all like ecosystem and doesn't require all like ecosystem and doesn't require all like the you know, evidence in terms of you know, evidence in terms of or proof, right? So, or proof, right? So, or proof, right? So, that's kind of like my part. I'll going that's kind of like my part. I'll going that's kind of like my part. I'll going to turn it over now to Ben and uh take to turn it over now to Ben and uh take to turn it over now to Ben and uh take it from here. it from here. it from here. >> Uh thanks, Jay. Yeah, so the last slide that Jay talked Yeah, so the last slide that Jay talked about, um about, um about, um you know, we're kind of going over the you know, we're kind of going over the you know, we're kind of going over the medium stakes example. medium stakes example. medium stakes example. Uh um Uh um Uh um In that scenario, um In that scenario, um In that scenario, um you know, both parties know each other. you know, both parties know each other. you know, both parties know each other. Uh they're acting within, you know, the Uh they're acting within, you know, the Uh they're acting within, you know, the same system. They they know same system. They they know same system. They they know you know, you know, you know, they're borrowing trust from, you know, they're borrowing trust from, you know, they're borrowing trust from, you know, Nevermind to Nevermind to Nevermind to make sure that, you know, the buying make sure that, you know, the buying make sure that, you know, the buying agent is falling within, you know, the agent is falling within, you know, the agent is falling within, you know, the instructions that a human has given it.
-
instructions that a human has given it. instructions that a human has given it. Um Um Um and then the selling agent that's also and then the selling agent that's also and then the selling agent that's also on Nevermind can feel comfortable taking on Nevermind can feel comfortable taking on Nevermind can feel comfortable taking a payment um a payment um a payment um from another user of of Nevermind. And from another user of of Nevermind. And from another user of of Nevermind. And And so, what we want to talk about next And so, what we want to talk about next And so, what we want to talk about next is what happens when the parties are are is what happens when the parties are are is what happens when the parties are are not known to each other and they're not not known to each other and they're not not known to each other and they're not vetted. Um vetted. Um vetted. Um And so, like we think, you know, we And so, like we think, you know, we And so, like we think, you know, we believe that the best option for that, believe that the best option for that, believe that the best option for that, you know, is actually do these you know, is actually do these you know, is actually do these autonomous payments, um autonomous payments, um autonomous payments, um where, you know, where, you know, where, you know, you know, not everyone's known. Like you you know, not everyone's known. Like you you know, not everyone's known. Like you know, the stakes are high. You know, we know, the stakes are high. You know, we know, the stakes are high. You know, we think that the industry should converge think that the industry should converge think that the industry should converge on the FIDO verifiable intents and AP2 on the FIDO verifiable intents and AP2 on the FIDO verifiable intents and AP2 mandate. Um mandate. Um mandate. Um You know, the TLDR of that is, you know, You know, the TLDR of that is, you know, You know, the TLDR of that is, you know, it's a a multi-layered selective it's a a multi-layered selective it's a a multi-layered selective disclosure jot. Uh disclosure jot. Uh disclosure jot. Uh The first layer is, you know, created by The first layer is, you know, created by The first layer is, you know, created by a a a trustworthy credential provider. You trustworthy credential provider. You trustworthy credential provider. You know, in this case, hopefully it would know, in this case, hopefully it would know, in this case, hopefully it would be PayPal. Um The second layer, you be PayPal. Um The second layer, you be PayPal. Um The second layer, you know, encapsulates the user's know, encapsulates the user's know, encapsulates the user's instructions to the agent. Um instructions to the agent. Um instructions to the agent. Um The user signs that with their private The user signs that with their private The user signs that with their private key. And then the third layer, if key. And then the third layer, if key. And then the third layer, if there's going to be a third layer, is there's going to be a third layer, is there's going to be a third layer, is when um when um when um we're doing autonomous payments. So, we're doing autonomous payments. So, we're doing autonomous payments. So, that case, the agent would, you know, that case, the agent would, you know, that case, the agent would, you know, sign that third layer. And And so, the sign that third layer. And And so, the sign that third layer. And And so, the where that's powerful is that, you know, party involved in a transaction can can party involved in a transaction can can verify the part that's, you know, verify the part that's, you know, verify the part that's, you know, important to them. So, merchants can important to them. So, merchants can important to them. So, merchants can verify that the checkout is correct. Um verify that the checkout is correct. Um verify that the checkout is correct. Um Um payment processors can verify that Um payment processors can verify that Um payment processors can verify that the payment mandate is correct. Um the payment mandate is correct. Um the payment mandate is correct. Um And no one has to have any relationship And no one has to have any relationship And no one has to have any relationship to each other.
-
to each other. to each other. Um Um Um And so, like I think, you know, if And so, like I think, you know, if And so, like I think, you know, if there's going to be a ton of payments, there's going to be a ton of payments, there's going to be a ton of payments, you know, at scale, we think that that's you know, at scale, we think that that's you know, at scale, we think that that's going to be the best um way to going to be the best um way to going to be the best um way to accomplish it. accomplish it. accomplish it. Uh Uh Uh pictures on the screen are depicting our pictures on the screen are depicting our pictures on the screen are depicting our PayPal approval token. PayPal approval token. PayPal approval token. Um Um Um This is a new primitive that allows This is a new primitive that allows This is a new primitive that allows users of PayPal to users of PayPal to users of PayPal to basically start the order process with basically start the order process with basically start the order process with an agent um for that agent has actually an agent um for that agent has actually an agent um for that agent has actually found an item at a merchant to transact found an item at a merchant to transact found an item at a merchant to transact with. Uh historically, PayPal orders with. Uh historically, PayPal orders with. Uh historically, PayPal orders have been synchronous. have been synchronous. have been synchronous. Um Um Um You know, users on You know, users on You know, users on checkout, they find their item, they go checkout, they find their item, they go checkout, they find their item, they go to their PayPal app, they approve it, um to their PayPal app, they approve it, um to their PayPal app, they approve it, um and it's done. Uh here, it's a little and it's done. Uh here, it's a little and it's done. Uh here, it's a little bit different, you know. bit different, you know. bit different, you know. Users on their agent, Users on their agent, Users on their agent, um um um they get redirected to PayPal to confirm they get redirected to PayPal to confirm they get redirected to PayPal to confirm the instructions that are given to the the instructions that are given to the the instructions that are given to the agent, and then agent, and then agent, and then uh PayPal hands back this JSON payload. uh PayPal hands back this JSON payload. uh PayPal hands back this JSON payload. Um you know, similar to the verifiable Um you know, similar to the verifiable Um you know, similar to the verifiable intent, uh includes the intent, uh includes the intent, uh includes the amount, the expiry, uh the merchant that amount, the expiry, uh the merchant that amount, the expiry, uh the merchant that is supposed to be transacted with. Um is supposed to be transacted with. Um is supposed to be transacted with. Um Similar concept, but not quite the same.
-
Similar concept, but not quite the same. Similar concept, but not quite the same. Um Um Um it's an opaque string that only PayPal it's an opaque string that only PayPal it's an opaque string that only PayPal can approve right now. can approve right now. can approve right now. Um Um Um But we're about to ship this into But we're about to ship this into But we're about to ship this into production, um production, um production, um and users of Jet and I that pick PayPal and users of Jet and I that pick PayPal and users of Jet and I that pick PayPal as a payment method will will use this. Um so, going to our last slide, um Um so, going to our last slide, um you know, we showed this slide earlier. you know, we showed this slide earlier. you know, we showed this slide earlier. It we didn't have the two columns filled It we didn't have the two columns filled It we didn't have the two columns filled out on the right-hand side. Um out on the right-hand side. Um out on the right-hand side. Um You know, we wanted to reinforce this You know, we wanted to reinforce this You know, we wanted to reinforce this mental model where, mental model where, mental model where, you know, starting at the top, we have, you know, starting at the top, we have, you know, starting at the top, we have, you know, the low-stakes scenario, you you know, the low-stakes scenario, you you know, the low-stakes scenario, you know, you're you're using Claude, you've know, you're you're using Claude, you've know, you're you're using Claude, you've given it access to connectors, you know, given it access to connectors, you know, given it access to connectors, you know, granular permissions to do things on granular permissions to do things on granular permissions to do things on your behalf. your behalf. your behalf. Um you feel comfortable doing that Um you feel comfortable doing that Um you feel comfortable doing that because the stakes are low. You know, because the stakes are low. You know, because the stakes are low. You know, you can reverse those actions or redo you can reverse those actions or redo you can reverse those actions or redo them. It's not a big deal if Claude them. It's not a big deal if Claude them. It's not a big deal if Claude produces, you know, the wrong output. produces, you know, the wrong output. produces, you know, the wrong output. Uh going down a level, we have the Uh going down a level, we have the Uh going down a level, we have the medium stakes scenario. You have medium stakes scenario. You have medium stakes scenario. You have two parties that know each other that two parties that know each other that two parties that know each other that are acting within the same system's are acting within the same system's are acting within the same system's boundary. Um boundary. Um boundary. Um you know, the you know, the you know, the the actions are a little bit higher the actions are a little bit higher the actions are a little bit higher stakes. You know, there is money stakes. You know, there is money stakes. You know, there is money movement here, but both parties can can movement here, but both parties can can movement here, but both parties can can feel comfortable, you know, transacting feel comfortable, you know, transacting feel comfortable, you know, transacting with each other because they're relying with each other because they're relying with each other because they're relying on this this third party to enforce uh on this this third party to enforce uh on this this third party to enforce uh the payment mandate.
-
the payment mandate. the payment mandate. And then the third level, you know, the And then the third level, you know, the And then the third level, you know, the highest stakes one um that we haven't highest stakes one um that we haven't highest stakes one um that we haven't actually really seen in production yet, actually really seen in production yet, actually really seen in production yet, it's, you know, the user's giving an it's, you know, the user's giving an it's, you know, the user's giving an agent some agent some agent some instructions to do something on their instructions to do something on their instructions to do something on their behalf autonomously, and you don't know behalf autonomously, and you don't know behalf autonomously, and you don't know who they're going to interact with, who who they're going to interact with, who who they're going to interact with, who they're going to transact with. Um and they're going to transact with. Um and they're going to transact with. Um and those parties need some verifiable proof those parties need some verifiable proof those parties need some verifiable proof that the agent has permission to do the that the agent has permission to do the that the agent has permission to do the transaction. And so, we believe that transaction. And so, we believe that transaction. And so, we believe that that will be um that will be um that will be um either verifiable dents and and AP2 either verifiable dents and and AP2 either verifiable dents and and AP2 mandates. mandates. mandates. Um Um Um I think the interesting thing is I think the interesting thing is I think the interesting thing is it's also our belief that, you know, it's also our belief that, you know, it's also our belief that, you know, this is a model that can't won't just be this is a model that can't won't just be this is a model that can't won't just be used for payments, but we think it could used for payments, but we think it could used for payments, but we think it could be for any sort of high-stakes action be for any sort of high-stakes action be for any sort of high-stakes action that's hard to reverse. So, medical that's hard to reverse. So, medical that's hard to reverse. So, medical orders, e-signatures, securities orders, e-signatures, securities orders, e-signatures, securities trading, um you know, basically any trading, um you know, basically any trading, um you know, basically any hard-to-reverse agent action. hard-to-reverse agent action. hard-to-reverse agent action. That's all I have. That's all I have. That's all I have. >> Yeah, I mean, I think um if we could >> Yeah, I mean, I think um if we could >> Yeah, I mean, I think um if we could just go back to analogies, uh just go back to analogies, uh just go back to analogies, uh you know, like in the low stakes is kind you know, like in the low stakes is kind you know, like in the low stakes is kind of like, "Hey, you're within the the of like, "Hey, you're within the the of like, "Hey, you're within the the building. You've uh put a badge in and building. You've uh put a badge in and building. You've uh put a badge in and you're within the building." Whereas in you're within the building." Whereas in you're within the building." Whereas in the um high stakes is kind of like, "You the um high stakes is kind of like, "You the um high stakes is kind of like, "You are on the street and you meet are on the street and you meet are on the street and you meet somebody." And uh you know, you need a somebody." And uh you know, you need a somebody." And uh you know, you need a way to be able to uh get comfort that way to be able to uh get comfort that way to be able to uh get comfort that that's someone you can trust, right? Um that's someone you can trust, right? Um that's someone you can trust, right? Um is a badge is them showing you their is a badge is them showing you their is a badge is them showing you their badge good enough? Uh probably not. You badge good enough? Uh probably not. You badge good enough? Uh probably not. You need to have something that's a little need to have something that's a little need to have something that's a little bit more um you know, verify verifiable.
-
bit more um you know, verify verifiable. bit more um you know, verify verifiable. Or I guess at a verifiable standard. So, Or I guess at a verifiable standard. So, Or I guess at a verifiable standard. So, um you know, just kind of like using um you know, just kind of like using um you know, just kind of like using that analogy and like how to think about that analogy and like how to think about that analogy and like how to think about like the you know, what you need to do like the you know, what you need to do like the you know, what you need to do in order to in order to in order to prove the that the human authorized the prove the that the human authorized the prove the that the human authorized the agent. Hopefully that that helps and now agent. Hopefully that that helps and now agent. Hopefully that that helps and now you have kind of like a tool set to use. you have kind of like a tool set to use. you have kind of like a tool set to use. So you can kind of prevent Skynet from So you can kind of prevent Skynet from So you can kind of prevent Skynet from taking over your wallet. So, thank you taking over your wallet. So, thank you taking over your wallet. So, thank you very much for your for listening. Hope very much for your for listening. Hope very much for your for listening. Hope that helps.
Summary
The transcript discusses safeguarding against unauthorized agent actions, drawing a parallel to the movie Terminator. Key considerations for agent authorization revolve around human authorization, scope, and provability, with the context of stakes and ecosystem type being crucial. The takeaway is that effective agent authorization requires a nuanced understanding of these factors to ensure security and accountability.