The more access an agent has, the more powerful it becomes.
Consumers hand agents every login factor. Arkose Labs founder and CEO Kevin Gosschalk explains what happens when people give an AI agent access to everything. People are giving agents their phone notifications and inbox, so agents can complete SMS and email one-time passcodes (OTPs) by themselves, along with banking credentials and payment methods. Kevin says years of building new authentication methods are collapsing back into one system that has everything, and that if it's compromised there is a problem.
About the speaker
The best way of using this technology, from a user experience standpoint, is you give it access to everything.
It can just autonomously satisfy email and SMS OTP.
So we've spent all these years building new authentication methods, but now unfortunately it's all being collapsed back to one system which has everything.
The best way of using this technology, from a user experience standpoint, is you give it access to everything.
The more access it has, the more powerful it becomes. I'm not necessarily advocating that from a security standpoint. Obviously there's a lot of asterisks around the idea of giving away credentials to a third party. People are doing this, so we should expect this as an industry. People are giving it access to their phone notifications, which means it can do SMS OTP by itself. People are giving it access to their inbox, so it can read emails and also do email OTP by itself. It doesn't need to check in with you.
It can just autonomously satisfy email and SMS OTP. People are giving it access to banking credentials, so it can log into a bank account by itself. People are giving it access to payment methods and all kinds of things. So we've spent all these years building new authentication methods, but now unfortunately it's all being collapsed back to one system which has everything. And if that one system gets compromised, then we're going to have a problem.