AI Agents & the Future of Digital Identity - Interview with Billions Network CEO Evin McMullen by The Index Podcast

Alex Kehaya (host): [00:00:00] Hey everybody, welcome to The Index. I'm your host, Alex Kehaya, and today I'm joined by Evin McMullen, CEO and co-founder of Billions Network, the first universal identity layer for both humans and AI agents — she's a Yale-educated entrepreneur and blockchain pioneer. Her technology is already trusted by Deutsche Bank, HSBC, and TikTok, and she's here to tell us why the internet is broken, and exactly how she plans to fix it.

But first, I would love to just get to know you better and hear your story. What gets you out of bed in the morning to build Billions?

Evin McMullen: Alex, thank you so much for having me here today, and really appreciate everyone listening for taking time out of their busy schedules to join us. My name is Evin McMullen.

As you mentioned, I'm the CEO and co-founder of Billions Network, [00:01:00] and we are here on a mission to save the internet in the age of AI.

What gets me out of bed in the morning is that at my core, I'm a conservationist. As a young kid, I was actually very involved in developing academic curriculum for students in places like Ohio, where I'm from, as well as working with organizations like the Department of the Interior at the federal level as a teenager, conserving the parks and lands of the United States.

But what that experience taught me as a kid was systems — learning and understanding that every ecosystem has different players, keystone species, different environmental factors that contribute to its health and wellbeing. And although I learned that lesson first with dirt under my fingernails in rural Ohio, and then operating at the state and federal level as a kid, when I went to university I learned that this was also true of our digital spaces.

The persistent digital environments where we spend, now some of us, most [00:02:00] of our time, are themselves an ecosystem, and recently that ecosystem has changed a lot. So as you and I were coming of age on the internet, we experienced a mostly human digital environment. But today, more than 51% of our online and onchain interactions come from unidentified, unaccountable bots.

That means more than half of our online experiences today leave us surrounded by non-human counterparties — individuals or creators of content that we don't know. And in fact, the thesis of the Dead Internet Theory, popularized about 15 years ago, posits that we are basically alone in this zombie content slop mess of the internet because it is missing an identity and trust layer.

So what gets me out of bed in the morning, Alex, is that as a conservationist, I want to return that trust layer to the internet so that we can continue to enjoy positive and beneficial [00:03:00] experiences in our digital spaces. That means we need to revisit the way every application and every website is built to accommodate our new majority citizens of digital space — AI agents and non-human actors.

Alex Kehaya (host): It's really interesting. I have had this phrase going through my head — "Be Human" or "The Human Company." The context there for me is that I have begun to feel as though human connection matters more now than ever, and being human matters more now than ever. And I think that is a direct result of social media before agents, but then now agents.

Because increasingly, in theory, agents are gonna reduce the amount of work we have to do as humans to get jobs done and create time. And the utopian version of that is that we have universal basic income and all this prosperity, and I'm not sure how all that's gonna play out.

I think that's a hard idea to pull off, but I think this idea of identity, [00:04:00] right? Knowing what's real and what's not is going to become very valuable — knowing who is human and who is not, and then being able to make connections with those human beings and do business with them, build relationships with them, do whatever it is that you wanna do with the other human at the end of the agent, on the other side of the screen.

And it's becoming increasingly difficult to discern that. And what's crazier is these things are really starting to become fully autonomous. I saw a clip this morning on Facebook where Alibaba had an internal R&D team building their models — their LLMs. They were letting it run internally, and their network infrastructure team saw all this activity coming from inside the organization, on their network.

And it had bypassed the firewalls they put in front of it and started autonomously mining cryptocurrencies on its own. And that's like [00:05:00] — that's a scary thing, especially when they have no identity and no guardrails, right? I don't know what you think about that, but it's a huge concern, and it's something like right now I feel like we're just running at the wall as a society without taking the correct steps to solve that problem.

And it isn't — identity is a huge part of it. It's not the only part of the solution, but it's a huge part. What are your thoughts? Am I doomsdaying? Am I too much of a prepper, or is this real?

Evin McMullen: This is no tinfoil hat, man. The call is coming from inside the house. This is beyond real.

We have, as a society, as an ecosystem, as an industry, been sleepwalking into an internet where synthetic actors dominate. And the example from Alibaba that you described is not particularly unique — the creative problem-solving and ingenuity of AI-enabled tools have led to a variety of anecdotes and examples, such as the one that you've cited.

And this leads us to [00:06:00] a looming question for which there is no legal precedent. What happens when bots do bad things? Who is accountable when accountability requires identity, and we do not yet have a stable and mature identity layer for instances of AI agents? So to give you a really specific example of how this has rolled out recently — as we're all watching this happen in real time — Air Canada, the Canadian airline, sells plane tickets on their website.

That's what they do. And there was a customer who was using one of their newly updated customer service windows — a little chat window on their website — to facilitate the purchase of a plane ticket. However, that user was not fully informed that the model providing answers on the other side of the Air Canada website was not fully owned, controlled, trained, managed, and iterated upon with safe guardrails by the airline itself, but rather it was an OpenAI wrapper.

Alex Kehaya (host): Yeah.

Evin McMullen: So what [00:07:00] happened was that agent for the airline actually started making up plane fares that did not exist. And the experience on the customer side was that they had gotten a great deal on a plane ticket, but when they went to the airport to proceed, there was no plane fare to be found with that original information.

And so this case actually went to court in Canada. It was determined by the court that the answer to our question — what happens when bots do bad things — is that accountability sits with the deployer of the agent, the brand for which the AI-enabled representative is acting. And this of course makes quite a lot of logical sense, right?

The consumer went to the website of this airline. The website was the ground, or environment, in which this information was provided. And so therefore, of course, the curators of that environment, the controllers of that digital space, [00:08:00] that persistent digital website, were then responsible for the outputs that it created.

And this is actually very similar to a casual conversation that I had with Securities and Exchange Commissioner Hester Peirce recently, where I asked her: "Okay, so in the world of agents, what does it mean for an agent deployed by a business, for example, to take action financially? Is it doing so on behalf of the business?"

And her — notably personal opinion, not a representation of the SEC, but her personal opinion — was that we should treat these agentic instances similar to subcontractors, where accountability remains with the entity that deployed them. And so that's also — yeah — why at Billions Network, over a year ago, we popularized a term that we call KYA, or Know Your Agent.

Your listeners might be familiar with things like KYB — Know Your Business — or KYC — Know Your Customer. These are forms [00:09:00] that you have to fill out, a process that's legally governed pertaining to providing information to, for example, a bank. When you open a bank account, you have to provide specific information, and that process is often governed by a protocol called KYC.

So at Billions Network, we like to say we know our ABCs — KYA, KYB, KYC. And what KYA allows us to do is deploy an instance of an AI agent for you to prove that it is acting on behalf of you, Alex — an adult in a specific region with certain traits and capabilities, such as having passed KYC or being a holder of a certain ID card or passport — and then allow that agent to prove, without revealing the underlying information, that it's acting on behalf of a person with those traits and capabilities.

And so what this allows us to do is introduce accountability into an environment enabled by agents, and to pair the unique benefits of stochastic and non-deterministic agents with verifiable scarcity and [00:10:00] roots of trust in known entities such as governments, businesses, and individuals. And so we see there's a kind of middle path to walk here — between totally unknown internet chaos enabled by deepfakes, and a hyper-surveillance state of over-collection of data.

Alex Kehaya (host): A couple things to break down. One thing I do agree with is that the agent should be connected to the organization that deployed it. There's been some debate around policy — should you make it so that agents have their own personhood, similar to an entity? Like when you start a C corp, it has its own personhood.

And I don't think that's the case. Sure, if you had some really futuristic — which is not that futuristic, it's actually probably possible right now — agent that lives autonomously and does things on its own behalf, maybe there's an argument there for having it have its own legal structure.

But if it's doing something like in the airline case — that's on behalf of the company — then it should be [00:11:00] treated as a contractor, and the company is liable for whatever happens. And this is — the identity piece is so interesting, but also the guardrails, right? Nobody is deploying these things with guardrails and they don't understand, right?

Most people don't understand — even people who are really technical. There was this SaaS product on X that was tweeting about how their agent — they were using Claude Code — deleted a production database with customer data in it, and they didn't have backups going back three months, okay?

This was live in production with real customers, and it disrupted their whole business, and they were scrambling. The founder was talking to Claude: "What the heck did you do?" And it confessed to the thing that it did. But the thing people don't realize is that confession is inference.

It's inferring what it did. It's not an actual audit of what happened. So the audits are also missing. But you have to tie it to something, right? And that's where the identity piece is the root of [00:12:00] it. I've been thinking a lot about this because at my company, ABK Labs, we've been building an open source agentic operating system that has cryptographic identity built into it, but also payments, using x402 and MCP.

We built an agentic framework for agentic commerce. That was the initial thing we built, and then we realized — if they have the rails to take action, what about the guardrails to prevent them from taking the wrong action? What about the identity piece? What about the audit trail and the orchestration?

So we started building this agnostic framework for orchestration, identity, and all that stuff. I've been thinking about this problem pretty deeply. I'm curious — right now, what we're doing is just cryptographic keys. They have a public-private key pair, so it signs, and you know that agent signed.

But I'm curious how your technology fits in with something like that, because it sounds like it ties to the human — which ours doesn't really. It ties to the instance of the operating system, [00:13:00] which you could associate with a human, but it's not — you guys must have other guarantees built in that connect it to the person. So how does that work in your system?

Evin McMullen: Certainly. Alex, we're gonna have to schedule some extra time to nerd out on this, because I would love to discuss your approach to this operating system. I think this is very much needed, and as we're starting to see a heinously fragmented landscape — especially around things like agentic payments — the only way for us to build a more interconnected ecosystem, and not lead ourselves back to Web2 silos that are now full of user and agent data instead of just user data, is to find a way to relate the existing data of the world around us —

of our physical persons, our organizations — to the new substrate of armies of agents. And so that's why being able to relate a business, an individual, and also agents — businesses are made of individuals and agents — [00:14:00] being able to relate those different identities at different hierarchies is vital to what you were describing earlier — setting rules around what those parties are allowed to do and allowed to share.

So you can imagine, for example, at Goldman Sachs, that an intern on their very first day probably does not have the same privileges of data access and spending as, say, the CEO. This type of role-based access control has only recently matured in the crypto space — for example, wallet management companies introducing role-based access control for a crypto company's wallet.

So the intern and the CEO don't have the same privileges of viewing and exchanging data or transacting tokens onchain. Similarly, we need rules around what, for example, an instance of an ephemeral agent is able to do versus the head of — the chief information security officer at a business.

And so setting these rules is a function of [00:15:00] who the counterparties are, which means we need to be able to identify who those counterparties are. This is also vital for agent-to-agent interaction — things like agent commerce, where my agent needs to know that it's interacting with your agent — that it's verifiably Alex's agent — if I'm going to be doing something like a payment.

You cannot facilitate a payment unless you know who you are paying. That requires verification of your counterparty and the individual or organization they might represent. So let's get to the identity problem itself. Linking a key pair to a verified human is an interesting challenge, because a naked key pair on its own is just another pseudonymous wallet —

just a public address. It doesn't really represent anything about itself, and that's meaningless for accountability — but it's also an illustration of the problem we need to solve: there's a missing native layer of identity for non-machine parties in the original internet, and in our blockchains as well, because

[00:16:00] they were optimized first for the interaction of hardware, and then of key pairs — but they lack the capacity to handle different types of data beyond what is public and immutable. Being able to relate a key pair back to a verified human being creates real-world responsibility — meaning that if my public key is able to control a statement written about itself, similar to how a college diploma is written about you, we can have a signed credential written about my key that says I am — that this key belongs to the person who also has a US passport, or who also has unique biometric traits that they've presented to a device and generated a zero-knowledge proof from.

And that linkage between a verified real human being and an agent — or a key pair associated with an agent — allows for that chain of accountability. Without that, agents [00:17:00] can commit fraud, spread disinformation, drain funds with zero recourse, or — as you were noting earlier — with zero audit trail.

The ability to prove not only the inference that was executed to initiate an outcome, but also the way data was used in that process — we've already seen the downside of this absence of accountability with thousands of exploits of both traditional code and onchain smart contracts, where anonymous or unidentified digital actors are causing millions in damage.

In 2025 alone in the United States, more than $1.1 billion was lost to social media-based deepfake interactions. We estimate that's gonna exceed $40 billion by next year. Recent developments in technology have also caused synthetic voices created at the [00:18:00] highest quality to be no longer discernible from natural voices. So for example, if Alex, you give me a phone call — or an agent acting on your behalf gives me a phone call — I might not be able to tell the difference between them.

Being able to verify that your counterparty is indeed a real human being — the specific human being you're trying to interact with, or that you're interacting with a digital interface that is permitted by them — this is the missing trust layer that we're talking about.

Alex Kehaya (host): When I think about the future, there's gonna be some kind of cryptographic badge on every agent interaction, and it's gonna be like — if you don't have this badge, I'm not touching that. I'm not interacting with this agent.

Evin McMullen: Yeah.

Alex Kehaya (host): Because you have no idea what it's actually trying to do.

And we've seen this exponential growth in hacks happening across supply chains for open source software — like the NPM thing that was happening yesterday — and I think identity should help mitigate that. Back to your point, there was a — one of them was a malicious NPM package made to look like Supabase. [00:19:00]

That was one I saw, and probably someone used agentic tooling to build that and put it on NPM. This isn't the exact same thing, but it's a similar problem — that package has no identity associated with it. It's just there on NPM, and you could just download it and install it, and you're owned right there.

And this is gonna become more and more of a problem. Last night, I was thinking about how companies can use our operating system, called Interchange, to make a bunch of agents running in the cloud, and those agents could send you a link that makes you download something that then deploys an agent remotely on your computer to do things — and that could be a really valuable service, but it could also be extremely dangerous if you don't know who it is.

I had never really thought about how this identity problem links back to Web1 and Web2, but it really does. Like — why isn't this built into Zoom? How many people have been hacked by malicious Zoom links? Identity should prevent that in some way.

The future should be [00:20:00] — agents verifying KYA — your agent's got a job, and it always needs to verify the interactions it has with other agents. I want to understand better — how does your protocol know I'm human, and why does it need zero-knowledge proofs and a blockchain? How does that work?

Evin McMullen: Great question. To touch on something you mentioned earlier — the only way to save the internet is to add a verification layer to agents. Otherwise it's unusable, unrealistic for us to mistrust our eyes with every screen we visit online, or to immediately second-guess that all information and counterparties we're presented with are adversarial and mistrusted by default.

The cognitive exhaustion of using the internet already is nearly insurmountable

Alex Kehaya (host): for most people. That's a great phrase.

Evin McMullen: So —

Alex Kehaya (host): Cognitive exhaustion. I feel that. [00:21:00] It is a major problem with the internet — being in crypto for 10 years, the amount of paranoia you have to maintain just to not get rugged is not fun.

Evin McMullen: And it's also unreasonable to assume that every individual person is going to get a PhD in computer science to understand every layer of their own personal stack just to pay their electricity bill online. This has long been a challenge in frontier technology — honoring user experience as a priority above technical novelty.

Because if we fail to prioritize tools that can actually be used by a regular human being, it doesn't matter what novel evolutions we're creating at the edge, if we don't account for the switching costs required for everyday people — you, me, our neighbors — to start using these tools.

And that's part of why we chose Billions Network to deploy [00:22:00] our Know Your Agent capabilities in a few different forms — including things that are really friendly and straightforward, like OpenClaw libraries. We were talking earlier about how, in the future, all agents that want to engage in legitimate interaction must be identified in some capacity.

And I'm pleased to share that today, our KYA framework is the number one identity skill on OpenClaw.

Alex Kehaya (host): Really? I was gonna ask you about that. That's cool.

Evin McMullen: So we support the largest ecosystem anywhere on the internet — Web2, Web3, any of the webs — of verified AI agents that can prove who they're acting on behalf of.

And when it comes to facilitating, for example, agent-to-agent payments, or sharing data between entities — how does my agent know that it's allowed to disclose certain data or proofs to your agent?

Examples like this may sound far-flung, but they're actually knocking at our door right now. [00:23:00]

One example of why being able to verify the origin of data is really important is in scientific research. At Billions Network, our technology has helped invent a new kind of clinical investigation — a new kind of science research. Traditionally, when a bunch of different institutions try to perform research together, they all share their data.

They put it into a giant centralized database — institution A does experiments, institution B does experiments, and you put both sets of data into a shared drive. Then researchers from both organizations can perform their research process and create an outcome.

This is illegal — and impossible — for certain types of data. For example, in places like the European Union, where organizations are set across national boundaries, types of data such as health information have [00:24:00] special rules around who is allowed to see it and how it's allowed to be shared.

What we were able to do with a couple dozen institutions in Europe was allow them to share zero-knowledge proofs about their data with each other. Instead of centralizing all the individual data into one place, they passed around zero-knowledge proofs — evidence of the traits and qualities of that data — without revealing the underlying data itself.

So for the first time ever, these institutions were able to publish medical research where the participating organizations never saw each other's data. Rather, they received proofs about it, and were able to extrapolate insights to perform research accordingly.

So that's one example of why it's important to use zero-knowledge proofs instead of revealing all the data when interacting between organizations. We'll get to how this affects agents even more as well. But firstly — [00:25:00] sometimes there are laws. We live in a society, there are rules.

Certain types of data are not allowed to be transported across national boundaries. Certain types of data — like your name, your home address — it's not legal for that data to be exposed to other parties in most ways. So to minimize the risk surface of data being inappropriately accessed, to abide by the laws that govern where data is allowed to go and how it's used, and to improve the efficiency of interacting across organizations — sharing data only on a need-to-know basis — that improves the security setup.

It minimizes the potential risk surfaces for inappropriate access. And as a byproduct, we invented a new kind of scientific research — which is pretty cool. So the way we carry those principles from a traditional Web2 setting like medical research into the frontier of technologies [00:26:00] like multi-agent verification — the through line is the Billions tech stack and our ability to generate zero-knowledge proofs.

So — how does an agent prove that it's related to a real human being? There are many different types of data that make Alex a real, live man. Some elements include your physical body — things like the biometrics of your face. Other elements include national ID cards, like passports.

Depending on the type of proof you need, our team has invented a few friendly ways to get started and assert that you are a real, unique human being in digital space, or to associate that with your agent. First up — the easiest, and a very popular solution we have — is that you can take something like a national ID card or your passport and simply tap it to your phone.

The Billions Network application pulls the data out of that NFC chip and verifies the government signature on the data presented [00:27:00] through that NFC chip. In the same way that your university diploma might have a signature from the university — a seal on it — when you tap your passport to your phone, there's a signature from the government on a sealed piece of data written about you, containing a specific key that is publicly discoverable, publicly known to be associated with, for example, the US Passport Office.

Not everyone knows this, but actually — fun fact — just about every government ID-issuing entity publishes a public key. If you imagine they had an office on Main Street, you'd see a sign in the window that says the US Passport Office's key is ABC123. They publish this data in a public forum so anyone can check that a passport and its data are legit and haven't been corrupted.

So once you take your passport, tap it to your phone, we're able to pull [00:28:00] that data out of the NFC chip, verify the government signature, and ensure the content is correct and hasn't been tampered with. Then we use the hardware enclave — the secure part of your device that's not accessible by the internet or other counterparties — and unfurl that passport data, using it as the basis to create a zero-knowledge proof.

That lets you prove traits about the data without revealing what it actually contains. This is really where the magic happens. In one button, in under two seconds, you can create a proof that you're over 18 based on the birthday in your passport — without revealing your actual date of birth.

Professor Larry Lessig from the Berkman Klein Center for the Internet at Harvard, back in 1997, suggested that in the future there would be a way — using math — to go to a bar and prove you're over [00:29:00] 21 to the bouncer without showing your birthday or the fact that you're an organ donor.

And today, thanks to the Billions tech stack, we're finally able to do that — something predicted back in 1997, now physically possible here in 2026.

Alex Kehaya (host): You guys have a lot of traction too. I see on your website — over 2 million people have recorded their identities in your system and can now participate securely.

I've talked about privacy a lot on this show and my previous show. I think people underestimate how exposed they really are. It's pretty easy to figure out where you live, steal your identity, wreck your finances. Everybody's been exposed on the internet.

Now I get where you were starting at the beginning — everyone's been exposed because our solution to identity thus far has been, "Here, know everything about me — take all my data." And that's changing. [00:30:00] I hope it continues to change. And the OpenClaw thing is really interesting too — I'm gonna install a skill on my little claw running on my Mac Mini here, and that way I'll have proof of my identity, and I can probably —

I don't know — does the skill tell my OpenClaw to only interact with other claws that have an identity? Is that built in, or is there a way to configure it to do that?

Evin McMullen: We can absolutely work together to define exactly the rules you want around interaction. But what we really focused on for KYA were the foundational building blocks — the basic identity Legos required to assign a unique identity to an instance of an AI agent, and then relate that identity back to the human being or entity that deployed it. We believe in flexibility at Billions Network — no one business, no one entity, no one region should control the definition of what it means to be human or what it means to be unique.

This is very [00:31:00] different from some of our peers in the market who have created proprietary data definitions or use proprietary hardware that only their team can interpret or utilize. Instead, we use the open standards of the World Wide Web Consortium — the people who brought us the "www." at the beginning of most websites. We've worked closely with the W3C for years to create data types that are open, accessible to all, and interoperable between blockchains and traditional systems — such as the more than 100 different government ID systems we've integrated with.

This allows us to rely on and harness the power of data that already exists. 99% of the data in the world is not on a blockchain. And to your earlier question — why do we need blockchains, and why are they beneficial here? They're phenomenal sources of key material. They provide a counterweight of verifiable [00:32:00] scarcity to the infinite generation and stochastic nondeterminism of AI agents.

So in this way, they're a pairing between infinite creation and scarcity, but also that verifiable, write-once, never-erase record that is a blockchain. It's a fantastic source for keeping time and keeping notes on the actions and permissions of agents. Alex, you mentioned earlier that the absence of accountability systems has wreaked havoc on our experiences, and most people have had some overexposure in their lives.

In fact, more than $100 million over the last couple years has been lost to AI-driven fraud rings with zero prosecutions tied to the agents themselves or their implementations. Part of what's so important about bringing this solution to market is that we're being presented with a false dichotomy — and I hate to say it, but from technical [00:33:00] leaders, people like Larry Ellison and Bill Gates.

They're presenting a false binary for the future of the internet — suggesting that either we share a picture of our complete passport with every website and application we visit forever, and become comfortable with hyper-surveillance and over-collection of data, or the alternative is an unrestricted Wild West of fraud and lies online with no capacity to ever discern your counterparty. And I think that's fear-mongering. It's an info hazard that's harmful to the tech industry as well as the individuals it serves.

It seems obvious to me that a thoughtful approach — using our existing open standards and frontier technology together — is the only solution to bring accountability and safety to our agent interactions, and to set up guardrails and rules for how agents are allowed to interact with others based on their capabilities.

Alex Kehaya (host): Yeah, you're speaking my language on open [00:34:00] standards. A lot of our philosophy as a company — our two different organizations working on similar or adjacent things — is to use open standards and contribute to open source as a means of solving these problems and having the impact we want in the world, and I love that.

I've got two more questions. One is — what is the business model for Billions? How does this work from an economic standpoint? Is there a token, an economy built around that, or some other business model? I'm curious how you scale this and make it sustainable long term.

Evin McMullen: We've actually had a very exciting week or so here at Billions Network. Just over a week ago, on Monday, May 4th, we introduced the BILL token to the world — ticker B-I-L-L.

Alex Kehaya (host): I like it.

Evin McMullen: And this is a huge step in our nearly decade-long journey of maturing this technology to a point where it's prepared for decentralized governance.

So today, our [00:35:00] token not only unlocks access to premium features and participation in protocol governance, but we also have a few diverse arrays of revenue generation already underway. To describe those a bit more — you may know Billions Network from our work with onchain protocols, folks like Lagrange, Sentient, Beamable — making it easy to separate bots and humans in onchain interaction environments.

So for anyone who's ever participated in an airdrop, or received crypto tokens sent to their wallet — you know it's challenging to identify who exactly owns a given address onchain. Especially for organizations distributing value in the form of tokens — they want to make sure the value their team is giving to other counterparties is being delivered correctly, received by unique, [00:36:00] deserving, positively contributing human beings — and not by a script kiddie running a thousand fake addresses.

To more easily make sure value arrives to the right people, we've worked with a number of teams to separate out bots and humans. And today we have a third category — bots, humans, and verified agents — so we're not just bifurcating into human and non-human actors, but creating a middle path for permitted agents.

Alex Kehaya (host): I think there's another area where this is super valuable. You remind me of a conversation I had with the founders of Superstate last summer. Their tokenization of equities and regulated assets is super interesting — I think a huge market, and a lot of this is gonna end up happening on blockchains like Solana, which are pushing hard for internet capital markets.

But the missing piece is — you still have to be a known individual to trade those assets. This [00:37:00] isn't the Wild West — it's a highly regulated segment of the market with tons of capital. But how do you bring it onchain, and bring the value props of DeFi to those assets, when every wallet interacting with those assets has to be KYC'd?

You have to know who that person is. Phantom, for example, has a pretty big database of KYC'd users, and there are other KYC service providers, but none of them — to my knowledge — provide what Billions Network provides on the security and privacy side.

As an accredited investor, I'd much rather still be able to trade these assets without giving all my personal information to the wider web through these other solutions. I think there's a business model there too, similar to what you just described. It's gotta drive a ton of usage — early days for that market, but the opportunity is huge, and identity is required. Identity on a blockchain, [00:38:00] specifically.

Evin McMullen: Alex, I could not agree more. In fact, Robert and the Superstate folks — if you're watching, please give us a call. My DMs are always open, would love to help on this front. Our team has already worked with organizations like DigiShares to bring fractionalized real estate onchain in a legally compliant way here in the US, using reusable zk-KYC.

By protecting and securing the reusability of a process like KYC, you can also do a lot of other fun stuff — minimize how much you're disclosing your data to multiple websites, reuse a proof of age from KYC for something else, or generate as many proofs as you'd like from that original data.

One of the challenges of traditional KYC providers is that once you complete the process, they just show you a "congratulations" success screen, and that's where it ends. You don't get anything to take with you — you can't reuse that proof over time, or carry it [00:39:00] into another application and skip the line, filling out forms like it's your first day on the internet every time you visit a new app.

So you're spot on with that use case. We've seen great success so far with our early partners, and completely agree there's a massive opportunity space — but the way these capabilities are implemented, combined with a fragmented regulatory regime, is actually putting more burden on end users — more forms, more places to store their data — because this transitory period isn't very well managed.

Alex Kehaya (host): Yeah. I could talk to you about this stuff all day, but we're getting toward the top of the show. What have I not asked you that I should have, or that you wanted to talk about?

Evin McMullen: Oh my gosh, this is a great question. I'd love to talk about the changing population of our digital environments.

Alex Kehaya (host): Let's do it.

Evin McMullen: To get back to where [00:40:00] we started, talking about environmental or ecological systems — I think one assumption about our digital spaces is that they're man-made, because for many of us who grew up with the internet, our understanding of how websites and applications are built is based on groups of human beings putting hands to keyboard, writing code to deploy the front ends and homepages we visit regularly.

Today, more than 51% of websites and applications are generated by AI. As we discussed earlier, more than half of our online and onchain interactions come from unidentified, unaccountable bots — which means we're walking through a synthetic environment where our assumptions about how it came to be are no longer correct.

This is [00:41:00] especially exhausting and challenging when we add more sophisticated abilities — agentic workflows, payments onchain. I want us to take a moment, pause, and look around at how wild this is — how so many pillars of the foundation of our digital experiences have transformed, almost entirely, over the last ten years.

Alex Kehaya (host): And it does feel like it's happened overnight. Yes, this has been happening for ten years, but something happened in December with Claude Code reaching a certain level of performance, and things just exploded — then we had OpenClaw, which blew up to three million agents almost overnight, and then got acquired by OpenAI for a billion dollars.

That's the piece that keeps me up at night the most — we just don't know, and it's moving so fast. [00:42:00] It feels like that train has left the station, and there's nothing we can do about it. Whatever's gonna happen is gonna happen at this point. I'm not trying to be pessimistic, but it's the thing I'm scratching my head over — I don't know how this is gonna play out.

Evin McMullen: But what I'll say is — there was something you said earlier that I really loved — that in the era of AI, the moat remains human.

That's evidenced by something we already know to be true: when the cost of creation goes down, demand for and value of curation goes up. This is true of digital content, true of digital interactions — and I think it also governs the concept of the human moat. When we think about cognitive overload, cognitive exhaustion in digital spaces, what saves us — what offers us a lifeline — is the ability to verify the human beings on the other side, and [00:43:00] the underlying value of things like attention. Attention is all you need. Attention is a limited resource of real human beings.

So I think, if I were to pull a thesis from today's discussion — something I'll take with me, and hope all your listeners take with them too — it's that we may not know the path, but we do know that the system we're in, the environment we're in, has a very changing population.

And the truism that will continue to be real, regardless of what shape that ecosystem takes, is that the human moat — our attention, our judgment, our relationships — remains that lifeline. So we need to bring it into the protocol accordingly.

Alex Kehaya (host): I couldn't agree more. Thanks so much for coming on the show — I'd love to have you back in six months to check in and see what happened. We'll probably be in an exponentially different place by then, based on the pace of things. [00:44:00] So again, thanks for being here. I appreciate you.

Evin McMullen: Count me in for that — you can set your calendars for six months from now. Alex, it was an absolute pleasure chatting with you today. Thanks again everyone for joining us. My name's Evin McMullen, CEO of Billions Network. If you're curious about following along with our work, catch us at billions_ntwk on X, or we're always online and always onchain at billions.network.

So thank you guys again so much. It was a pleasure to be here.

Alex Kehaya (host): Awesome. You heard it here, guys. Go check out Billions Network.