How to stop AI fraud? Ask your bank.

Hadi Partovi recently posted a video with a brilliant intro to ChatGPT explained by Mira Murati , who led the team that built it, and Cristóbal Valenzuela from Runway. Generative AI is going to add massive value to the world. The video also provides a clue to solving the Internet's biggest problem.

AI makes solving the problem more urgent because AI also gives superpowers to the forces of deceit and deception. Bad actors will use AI to defraud, misinform, hack, and harm us all. We will not be able to stop the AI bots because they will be so good, we won't always be able to detect them. As Yuval Noah Harari says, "If I can't tell the difference between AI and a human, that is the end of democracy, because that is the end of meaningful public conversation."

How do we preserve truth and trust on the Internet? The solution is to flip the problem and give humans the power to prove their identity and their humanity online. The best identity verification solutions always require widely trusted 3rd parties. Today, in the physical world, government ID credentials like driver's licenses serve this purpose.

Online the challenge of verifying identity becomes much harder. There are two parts to the problem.

1. How do we accurately authenticate the person on an anonymous session?
2. How do we preserve our privacy?

The best way to solve this is with Bank-based ID. Banks are not the only option, but banks have critical advantages. There are thousands of banks. People have lots of choices. Banks are well-regulated and required to protect client privacy. Banks will recognize that client identity credentials are property and banks are already in the business of managing assets and providing custody (aka vault) services. Banks have a business model that matches with the goal of protecting assets and protecting privacy. Tech companies generally do not. Instead, Google and Facebook treat the human and the human's attention as the product and make a profit by compromising privacy to sell taste and attention to advertisers.

Banks have another advantage which relates to the hint described by Cristóbal Valenzuela in the video. ChatGPT works because it picks the next token using extensive context. To accurately authenticate the person on an anonymous session, such as any person arriving on any app or any website, you have to include deep context about that person's behavior biometrics, pattern of network usage, device ID, time of use, and history of movement. Lex Fridman 's PhD thesis introduced this concept to the world, and in many ways, he is the giant upon whose shoulders the solution to the Internet's greatest problem will be built. To authenticate someone, we can't rely on just the device. It can be compromised. We can't rely on biometrics. They can be faked. But in combination with a lifetime pattern of behavior, a user's preferred bank is in the best position to say "The person on this website at this time really is who they claim to be."

Recently, I was speaking with George Damouny . He called what the banks can offer Continuous Authentication. It's a clever way to think about it. Until now, Continuous Authentication has been used by companies for their employees. In the spirit of filliping the problem and empowering users with the tools to prove who they are, Bank-based ID takes the idea of Continuous Authentication and helps people leverage it for their entire life.

I am who I claim to be. I can prove it to you. Just ask my bank. They have known me for years. They follow me everywhere. I am OK with them following me and continuously authenticating me because it helps protect my money. And I trust my bank to protect my privacy and only share the information I want them to.

With the power of Bank-based ID to protect privacy and provide ID verification based on government-regulated KYC and Continuous Authentication, we can leverage the power of AI while still having tools to establish truth and trust.

What is the Internet's greatest problem? The Internet's greatest problem is the missing Identity layer. Online, no one knows you are a dog and no one knows you are a bot. If don't know who is on a site, we can't build trust. Bad actors start by hiding their identities. The Internet makes it easy. The result is fraud, theft, hacking, misinformation, trolling, bullying, and time wasted building trust.

The solution is to give people the power to prove their identity. The best way to do that is with Bank-based Identity.

My company, IDPartner Systems is working to deliver a global Bank-based Identity Network. Verified and Trusted. Reusable Everywhere.