CAPTCHA was built to keep bots out. But what happens when we actually want AI agents on the internet?

I've been thinking about how the internet might change as AI agents become more autonomous. Back then, we built CAPTCHA to distinguish humans from bots and prevent automated abuse. It became a common part of the web. But now we're moving toward a world where AI agents can browse websites, use tools, make decisions, complete tasks, and even interact with other agents on behalf of their owners. And that got me thinking: are we still approaching bot verification the right way? Instead of simply asking "Are you human?", maybe websites will eventually need to ask: Are you a human or an AI agent? If you're an agent, who owns or operates you? Can you prove you're authorized to perform this action? Can your permissions be verified and revoked? I'm aware there are already technologies like OAuth, DIDs, Verifiable Credentials, and agent identity projects exploring parts of this. But I'm curious whether we'll eventually need something as widely adopted and easy to integrate as CAPTCHA, except designed to verify legitimate AI agents rather than block them. I'm not proposing a solution yet. I'm just interested in how developers think this should work. Do you think the internet needs a standardized verification layer for AI agents, or can existing authentication and authorization technologies already handle this?