Contact The Editor
Articles By This Author
SAN FRANCISCO — Clement Anthonioz, a 38-year-old telecommunications professional, walked into a sleek store in the Union Square shopping district Thursday and stared into a white orb as it scanned his iris to generate his "proof of personhood."
It was opening day for one of six U.S. storefronts promoting World, a digital identity system built on the idea that dating apps, government agencies and more urgently need a reliable way to distinguish people from machines.
"There's going to be more and more AI agents online - and sometimes you want to prove that you're a human," said Anthonioz as his beagle, Tixi, waited patiently. (There is not yet a proof-of-pet option.)
Anthonioz was willing to put his eyeball on the line for a World ID "proof of human" credential because he believes it will be useful in the future for all kinds of services. "I think at some point to vote we'll need this," he said. But some privacy experts question whether the World network will be able to deliver on its promises - or keep user data safe.
The orb that scanned Anthonioz's iris was made by start-up Tools for Humanity, which makes the technology behind the World network and was co-founded in 2019 by Alex Blania and Sam Altman, CEO of ChatGPT-maker OpenAI.
The company previously scanned irises outside the United States, including in Argentina and Kenya, where regulators in 2023 halted its operations for months to clarify its privacy and security practices.
As World launched in the United States, Tools for Humanity said that 12 million people have completed an iris scan. It also opened stores in Miami, Los Angeles, Austin, Atlanta and Nashville.
Enrollment involves the orb taking images of a person's face and then comparing their eyes to a selfie taken using the World app on their phone. The entire process takes place on-device, Tools for Humanity said, and each World ID is encrypted with an anonymized, cryptographic key.
Once the process is complete, the World app informs a user that their data is being deleted from the orb. A person can use the app with compatible services to prove that they are not a bot or an algorithm.
At a launch event at a warehouse space on San Francisco Bay on Wednesday night, attendees dined on tuna poke bowls and roasted marshmallows with Ghirardelli chocolate at a make-your-own-s'mores bar. Frozen margaritas and gin slushies were on hand to loosen any inhibitions before an eyeball scan.
A public relations representative confirmed that the disc jockeys spinning that night had also created World IDs. San Francisco Mayor Daniel Lurie (D) and California Gov. Gavin Newsom (D) were spotted at the festivities, although their offices said they did not create World IDs.
The company announced partnerships with Visa, dating app provider Match Group, video game hardware maker Razer and payments provider Stripe.
Altman, whose OpenAI has popularized use of text and image-generating AI tools, told the crowd of nearly 1,000 attendees that five years ago it became clear to him and Blania that "we needed some sort of way for identifying, authenticating humans … in a world where the internet was going to have lots of AI-driven content."
That conversation culminated in the creation of a "proof of human" system, Blania said, that is global, privacy-preserving and has a financial component.
Anyone who enrolls with an iris scan is rewarded with cryptocurrency called Worldcoin. Those who did so at the event Wednesday received 150 of the crypto tokens, each of which is worth a little over a dollar. One attendee said she wasn't thrilled with the idea of an eye scan, but that she'd do it for the free crypto.
Lorrie Cranor, who directs the CyLab Security and Privacy Institute at Carnegie Mellon University, saw Blania and Altman present the World vision at a privacy conference last week. Though the company positions its "proof of human" credential as a way of authenticating someone is a person and not a robot, she said it was not clear why having a scan of her iris was necessary or that it would accomplish that goal.
"Now I have features from the scan of my iris encrypted and stored on my phone, but if somebody else gets access to my phone or if a robot takes over my phone, does that mean that they can demonstrate that they're human or maybe even me?" Cranor said in a phone interview Thursday.
She said that she asked this question of the person accompanying the iris-scanning orb and wasn't entirely satisfied with the answer.
Cranor said her research has shown that people still care about privacy, but they've given up on protecting it tightly. "The cat is out of the bag," she said.
She has direct experience of the human-or-AI problem, but is unsure World can help. When doing research surveys, Cranor has often received answers she suspects were formulated by ChatGPT.
"We've been trying to come up with all sorts of ways of figuring out who is not really human taking our surveys. It's really hard. … I have trouble believing if this system will actually solve this problem."
Rachael Yong, a start-up founder, said she wasn't wild about someone scanning her iris when she first heard about World a few years ago, but at the Wednesday event was impressed by the partnerships announced. "I have Stripe. I have dating apps," Yong said. "It's interesting to see that they're trying to integrate [World ID] into people's lives."
Other attendees, like Chris Jackson, a marketing and advertising professional in San Francisco, were wooed by Altman's star power. "I genuinely think he has a vision of what the future is going to be," Jackson said.
He thinks World ID could be useful in verifying people's ages and personhood online. More than a dozen states have already passed or enacted laws requiring online age checks, with many companies already offering the technology to social media companies such as Meta and TikTok.
Debra Farber, privacy engineering manager at Lumin Digital, suggested in a recent LinkedIn post that data associated with a World ID such as where or how often it was used could be used to pierce a user's anonymity. "If World ID is linked to other accounts (e.g., social media, banking), it might be possible to re-identify individuals despite its privacy-preserving design," she said.
Farber also sees hypocrisy in the company's mission. "World attempts to solve a crisis of human authenticity that Sam Altman's own work at OpenAI is accelerating, creating a feedback loop where the harms of one system justify the surveillance of another," she wrote in a message to The Washington Post.
The software behind the World network and the iris-scanning orbs is open source, and Tools for Humanity has said it welcomes scrutiny from privacy and digital rights organizations.
Anthonioz said he trusts Tools for Humanity with his biometric data in part because it's a well-funded American company, and he's impressed the background of the team. Besides, he said with a shrug, he's already given other companies access to much of his biometric data.
"I think this is safer than Face ID," Anthonioz said, referring to Apple's method of unlocking and storing passwords on his iPhone.