This post originally appeared in the Packet Pushers’ Human Infrastructure newsletter, a weekly mailing of essays, links to technical blogs and IT news, and whatever else think is interesting. Subscribe for free here.
Let’s say I invent an autonomous mobile robot. It can lift heavy items for you, wash your dishes, do your grocery shopping, walk your dog, and handle many other tasks you find tedious. For a robot it’s surprisingly personable and charming. People buy a lot of them, and soon they’re everywhere. Even if you don’t own one yourself, you can’t avoid them.
For reasons no one clearly understands or can explain, sometimes these robots hit you on the head. There’s no warning and no signs of malfunction. A robot whacks you on your noggin and then goes back to whatever it was doing.
So far, no one’s been killed or seriously injured, and the blows don’t hurt much. But why do you have to be hit at all? And could it get worse?
As the inventor, I don’t want you to stop buying my robots. So I design a safety helmet. This is no mere bike helmet or hard hat. It’s sleek, high-tech, and specifically crafted to absorb robot blows. Because my design requires that I capture precise biometric measurements of your skull, no one but me can manufacture these helmets.
Also, you have to wear it all the time because you never know when a blow is coming. But isn’t that just a minor inconvenience compared to all the benefits my robots provide?
And because I’m a big thinker, I recognize that the helmet is not just a helmet–it’s a platform. I can add all kinds of interesting things to it: media devices, communication systems, digital currencies, and more.
When enough people have this headgear, we can network them to a global blockchain that will allow us to communicate, trade, and share in a prosperous helmet-based future!
In the meantime, I hope my platform idea distracts you from the fact that you had to buy the helmet because of all the hitting-you-on-the-head robots I invented.
On The Internet, No One Knows You’re An AI
This robot story popped into my head while I was reading about Worldcoin. It’s the latest venture of Sam Altman, one of the people behind OpenAI and ChatGPT. What is Worldcoin? It’s a proposal to create a “globally-inclusive identity and financial network.”
Worldcoin includes a blockchain and digital currency (of course), but its most ambitious goal is to develop a universal “proof of personhood” system so that we’ll be able to distinguish humans from AI. This proof of personhood requires having your iris scanned by a special camera called an Orb, developed by yet another Altman startup affiliated with Worldcoin. (Eye-scanning Orbs don’t sound creepy at all, right?)
The whitepaper explaining Worldcoin comes with all the de rigueur sentiments about openness and open source, and how this system will usher in a new dawn of shared benefits and prosperity. The mission statement says Worldcoin could, if successful, “considerably increase economic opportunity…enable global democratic processes, and show a potential path to AI-funded UBI.”
Those are laudable ambitions, but take a moment to let the big idea sink in: Sam Altman, who’s had immense success developing and selling generative AI tools that fluently mimic human communication, is now telling us that it’s necessary to sign on to this “proof of personhood” scheme so that we can tell real people from the fake ones he’s helped create.
The thing is, he might be right. Internet communications and social media platforms are already creaking under the weight of bots, sock puppets, spam, and other forms of fakery. And the fakery is improving at a frightening rate. Large Language Models (LLMs) that power tools such as ChatGPT have gotten very good at deceiving humans.
Throw in AI-powered deep fakes and synthetic media, which are already popping up to muddle the 2024 US presidential election, and it’s easy to imagine an online future where it’s impossible to tell what’s true, what’s false, or who’s even real.
Will This Time Be Different?
Worldcoin is bold and ambitious. It tackles very hard problems around identity while trying to preserve individual privacy. Maybe it’s found the right structure to create a decentralized network that could be globally owned. Maybe it really could be a mechanism for a more equitable distribution of resources.
And yes, we do need reliable mechanisms to parse the real from the artificial online. But does it need to be Altman’s system? Does it need to be wrapped up in a world-swallowing platform that bolts your identity to a currency and a tradiing network and who knows what else?
Take some time to read the Worldcoin whitepaper. Like an LLM, it’s pretty convincing.
Then ask yourself how many times we’ve fallen for a techo-utopian pitch that promises global connection and prosperity, but in the end just creates a handful of new billionaires and leaves the rest of us more polarized, confused, and wondering where the next blow will come from.