williamcotton.com

Authentically Human in a World of ChatGPT

February 7th, 2023

As outlined in The Tyranny of the Anonymous, anonymity on the internet is clearly beneficial in some cases and technically here to stay, but there are plenty of reasons for optional forums that do not allow anonymous participation

A rapidly accelerating use of anonymity is to hide the fact that a participant is not even human, let alone a troll or state-sponsored actor. Large-language models such as GPT will increasingly be used in ways to appear as though something published to a website was from a human author and for reasons that are too unpredictable to enumerate. It seems wise to offer a solution.

Relying on private companies such as Meta or Amazon for identity verification is unwise for a number of reasons, but most importantly that private companies are in the business of providing services to their customers. Just look how easy it is for someone to engage in social engineering when faced with a person trained in part to keep the customer happy! Can you imagine the "crying baby trick" working with the an employee of the DMV? From my experience the DMV is not in the business of keeping anyone happy and has plenty of experience and protocols in place to limit fraudulent identification.

So how do we go about such state-level identity verification?

From The Tyranny of the Anonymous:

In a practical sense, we're in luck, as the United States Department of Defense already uses a common authentication system known as a Common Access Card. This card is a photo ID that contains a private key that is used to not only login to DoD computer systems but also to sign encrypted messages between individuals.

There is no real technical challenge for each state to issue CAC-compatible IDs through their respective departments of motor vehicles. The technology is literally battle-tested, cheap and abundant.