Can “Proof of Human” protocols actually secure us against the AI surge?

Before dismissing some of these tools, World Coin, that were meant to verify human identity, I thought I'd do my own research on how they approach security in this new era of synthetic media. Hoping to get some critical feedback from you all on whether or not this is even a good approach or if we're just seeing more tech solutionism at work.

What I think is interesting in their architecture is that it’s non-custodial, meaning that your keys are stored locally on your device. More interesting is that it uses Zero-Knowledge Proofs to verify that you are indeed human without ever revealing your true identity. We're talking about a world where deepfakes and AI bots are being used to impersonate humans all around us. So, it’s time we had that conversation.

I think that, in comparison to traditional apps that store sensitive data in plain text or on central databases that AI bots could scrape with ease, we're talking about something that’s decentralized. Of course, I think that maybe the biggest threat is still human error, phishing, or even that we're creating something that’s being co-opted by what we're trying to avoid.

Does anyone have any insights on any security incidents that have occurred with tools that do something similar with “human verification”?