Worldcoin: a solution in search of its problem

Having my eyeballs scanned by a shiny chrome orb so I can someday receive cryptocurrency disbursements because artificial intelligence has stolen my job sounds like something from the pages of a half-baked sci-fi novel. It also sounds like the kind of operation that venture capitalists would value at over a billion dollars.

Worldcoin advertisements in San Francisco. QR codes have been covered with a quote from the Surveillance Technology Oversight Project (via @[email protected])

The premise is simple, they say: As artificial intelligence becomes more sophisticated, approaching the level of human-superior artificial general intelligence, it will both create wealth and disrupt labor markets as human workers are replaced by machines. It will also become increasingly challenging to distinguish human from machine, as the current-day problem of bots mimicking humans online is made worse by sophisticated AI fakes. Or at least, that’s today’s problem. Check back in a month or two and see if it’s changed.

Worldcoin is, at the moment, a project to distribute cryptocurrency tokens (also called Worldcoin, or WLD) to those who confirm they are human by having their irises scanned by a custom piece of hardware that both captures its subject’s unique iris “fingerprint” and performs biometric scans to ensure it’s scanning a living, breathing human being and not a printout or some other fake. That custom hardware just so happens to be a chrome orb that evokes HAL 9000.

HAL 9000 (left), Worldcoin Orb (right)

Worldcoin was founded by Sam Altman, the “tech visionary” du jour who is behind OpenAI. That’s right, the guy who’s going to sell us all the solution to a worsening AI-powered bot infestation of the Internet and to AI-induced mass unemployment is the same guy who’s making the AI in question.

Sam Altman (TechCrunch, CC-BY 2.0)

He is selling the antidote to the poison he is, coincidentally, also selling.

The identity problem

In June 2022 I wrote an essay titled “Is ‘acceptably non-dystopian’ self-sovereign identity even possible?“. This was prompted not by Worldcoin, but by a different identity-related project: Vitalik Buterin’s conception of “soulbound tokens”. In a May 2022 paper he co-authored with E. Glen Weyl and Puja Ohlhaver, he wrote that these identity projects (and related ideas they group under the umbrella of “decentralized society”) must meet the rather low bar of being merely “acceptably non-dystopian” in order to be worth pursuing.

In that essay, I wrote about the problem of decentralized identity: that is, how do you determine that someone is who they say they are without relying on a centralized authority (e.g., government-issued identification)? This has been a popular topic in the cryptoverse because of the Sybil problem: the challenge of ensuring that one individual does not operate multiple identities (or crypto wallets) while also respecting anonymity (or pseudonymity). Worldcoin is among the projects trying to solve this problem, sometimes termed “proof of personhood”, but hardly the only one. Proof of Humanity, BrightID, and others are doing so as well. They use a range of approaches, from biometrics (Worldcoin) to web-of-trust vouching (BrightID) to some sort of a mix (Proof of Humanity incorporates both vouching and uploading a video of one’s face for verification).

Identity projects aim to answer one or several of the following questions:

Is this user a human?
Is this user a unique human? (i.e., do they only control one identity in a given network?)
Can this user prove they meet some criteria? (e.g., are they over 18? are they a U.S. citizen?)
Can this user prove they are a specific person? (e.g., does Molly White control this identity?)

If you’ve ever solved a CAPTCHA you probably understand why it’s useful to answer the first question — bot prevention is important and inarguably becoming more important as bots become more convincing, more disruptive, and more capable of evading anti-bot measures.

The second is useful for ensuring fairness in systems such as voting. In the cryptocurrency world, many DAOs follow the one-token-one-vote model, which makes them susceptible to control by the wealthiest actors. A robust proof-of-personhood solution could, ideally, allow for one-person-one-vote without sacrificing crypto’s beloved pseudonymity. Some also envision decentralized proof of uniqueness as helping with fraud protection in other systems, ranging from ensuring people only get one NFT in an NFT airdrop all the way to critical social welfare programs or universal basic income schemes.

The third question is distinct from the fourth in that there are times when people might wish to provably answer specific yes or no questions such as “are you over 18?” or “are you a U.S. citizen?” without disclosing their full identity. This is not widely done today: websites generally either ask if you meet the criteria (and simply trust that you aren’t lying), or require you to submit a government-issued ID to prove it (and thus also require you to disclose your full identity to them).

The fourth question is critical for high-importance activities, like signing legal documents, opening a bank account, or applying for a loan.

Worldcoin’s goals

Worldcoin is primarily concerned with answering questions 1 and 2: ensuring everyone who is represented in the Worldcoin network is a real person, and only controls one identity in the system. The reasons why they want to do this have shifted since the project emerged in 2019, making the project hard to pin down. First they seemed most focused on wealth redistribution. Then, during the height of crypto hype, the story centered around onboarding new users to crypto and solving its Sybil problem. More recently, Worldcoin pivoted to the more AI-focused story now that AI is the hot big thing and founder Sam Altman has become its poster boy.

The types of things Worldcoin says it could one day do are lofty: “considerably increase economic opportunity, scale a reliable solution for distinguishing humans from AI online while preserving privacy, enable global democratic processes, and show a potential path to AI-funded UBI [universal basic income].“

That’s right: the founders aren’t looking to merely create the next generation of CAPTCHAs, they want to form the base of future democracy and social welfare.

Mechanics

To join the Worldcoin network, people download the World App crypto wallet on their smartphone. They then find a nearby Orb, submit to its iris scan and other biometric humanness-detection, become “Orb-verified” and receive a World ID. To accomplish this, the Orb scans the iris, applies its special algorithm so that it can compare the iris scan to the others in its database and ensure uniqueness (while accounting for the fact that two scans of the same iris may not be visually identical due to factors such as lighting or angle), and verifies the ID if the iris is indeed new to the database.

World App (via Worldcoin)

Worldcoin is very quick to insist that they do not store the iris scan data directly, but rather store an “IrisCode”, which they describe as a mere “set of numbers” on its website.

2 The IrisCode has been widely described in media as a “hashed” version of the iris scan, and indeed used to be called an “IrisHash” by WorldCoin, but references to hashes seem to have been (somewhat incompletely) scrubbed from the website as of late. Worldcoin tries to insist that they don’t store sensitive biometrics, a claim that requires everyone to simply go along with their assumption that a per-person unique IrisCodeitselfis not sensitive data. It’s also not terribly clear yet what kind of data might be leaked by the IrisCode — for example, Vitalik Buterin has questioned if some traits captured in the code might reflect things like sex, ethnicity, or medical conditions.3

Worldcoin is also considerably less forward about the fact that they encourage users who sign up to “opt in” to image custody. For those who opt in, WorldCoin continues to store the original iris images, they say “because the algorithm that computes the iris code is still evolving to make sure it can support signing up everyone”.

4With Orbs still relatively scarce, users face the risk of being removed from the pool of verified World ID holders as the algorithm is refined, unless they either have ongoing access to an Orb at which they can be re-scanned, or acquiesce to their original data being retained.

Because the vague promises of maybe someday enabling DAO voting or AI-necessitated UBI are both intangible and probably unappealing to many of the massive number of people Worldcoin hopes to onboard (ranging from several billion to every single human on the planet, depending on which exec you ask and when), Worldcoin has decided to just pay people for their eyeballs.

But rather than handing out cold hard cash, those who sign up receive 25 Worldcoin tokens (WLD), and the opportunity to claim 1 WLD per week going forward. Or at least those in approved jurisdictions do — US-based users can’t receive tokens in return due to pesky regulatory concerns, and in several states or cities can’t be scanned at all.

5The app is also not available in some jurisdictions, including mainland China.6And privacy regulators are already sniffing around in various European markets where Worldcoin has recently started scanning irises in a push following its big launch this week.7

At the beginning of Worldcoin’s iris-scanning endeavours, the WLD that people received was no more than an IOU, since the token hadn’t yet launched. Since the token launch on July 24, the price has fluctuated between $1.94 and $2.69 — as of writing, it is hovering at around $2.35, making the initial 25-token distribution worth around $59 to anyone who immediately cashes out.

What could possibly go wrong?

If by now you’ve found yourself thinking “scan my irises and give the data to a bunch of VC-backed tech bros in exchange for tokens that may or may not be worth around $60? sign me up!”, well, keep reading.

A caveat: Worldcoin is at this stage so incredibly vague about what exactly they envision people using the project for that it is difficult to analyze. I would certainly be asking very different questions of a project that simply aims to ensure people are only receiving their fair allotment of promotional NFTs than of one with aspirations of becoming the voting apparatus for “global democracy” or the operator of a worldwide universal basic income program.

A shaky ethical foundation

Before launching in to future-facing issues with Worldcoin, it’s worth touching on its history a little bit. In April 2022, MIT Technology Review and BuzzFeed News nearly simultaneously published longform articles stemming from their investigations of the project, particularly focusing on their experimentation in low-income communities, often in developing countries, and on individuals who did not always understand what they were agreeing to. The articles detailed numerous issues with the company, including unconscionable treatment of their hired “Orb operators”, the widespread use of questionable tactics to entice new people to sign up, inconsistent messaging about exactly what kind of data was being collected or preserved, and lack of compliance with local data privacy policies. Both articles are well worth the read.

Worldcoin testing in Indonesia (via MIT Technology Review)

The premise itself

Much of Worldcoin’s promises are predicated on the questionable idea that highly sophisticated artificial intelligence, even artificial general intelligence, is right around the corner. It also hinges on the “robots will take our jobs!” panic — a staple of the last couple centuries — finally coming to bear. Worldcoin offers other use cases for its product too, like DAO voting, but it is not the promise to solve DAO voting that earned them a multi-billion dollar valuation from venture capitalists.

Other use cases that Worldcoin has offered seem to assume that various entities — governments, software companies, etc. — would actually want to use the Worldcoin system. This seems highly dubious to me, particularly given that many governments have established identification systems that already enjoy widespread use. Some even employ biometrics of their own, like India’s Aadhaar. There’s also the scalability question: Worldcoin operates on the Optimism Ethereum layer-2 blockchain, a much speedier alternative to the layer-1 Ethereum chain to be sure, but any blockchain is liable to be a poor candidate for handling the kind of volume demanded by a multi-billion user system processing everyday transactions.

And finally, bafflingly, Worldcoin seems to think it — a VC-backed corporation — is best positioned to save the world from this forecasted AI-induced economic upheaval via “AI-funded universal basic income”.

Not-so-decentralized identity

If you ask the proof-of-personhood folks, centralized identity systems suffer from unacceptable flaws, namely lack of privacy and the risk that the maintainer of the identity system could act maliciously towards members of the network (or could be corrupted or taken over by someone who does). They have some good points.

But Worldcoin itself is enormously centralized, and at this point, talk of decentralization is little more than handwavy promises. The custom Orb hardware presents a massive obstacle to decentralization that Worldcoin doesn’t seem to have meaningfully grappled with. If Worldcoin is the only group producing these Orbs, they exercise sole control over them — which, in turn, provides them the sole ability to introduce backdoors.

If the Orbs hardware is “decentralized”, which Worldcoin says they intend to do,

9they then have to ensure that the Orbs are properly constructed following the design specifications they’ve provided, and haven’t been modified to maliciously create IDs outside of the intended mechanism. Worldcoin speaks vaguely of a third-party auditing system and allowlisting process that would attempt to catch any such malicious Orbs, but the scale of such audits required to allow for the quantity of Orbs to achieve billions of signups would be enormous. Furthermore, because Worldcoin incorporates cryptocurrency distributions, any bad actor who was able to slip through a malicious Orb capable of generating fake IDs could rapidly siphon WLD tokens, and even if discovered, the past distribution could not be reversed — they could only be prevented from continuing to create new IDs.

The cryptocurrency industry is rife with projects that embrace the idea of “progressive decentralization”: beginning out as a highly centralized project run by a small group, but promising to eventually turn over control of the project to a DAO. Few ever follow through,

10but it is a convenient way to stave off criticism.

Privacy

When questioned about the wisdom of attempting to form a huge database of iris scans, Worldcoin argues that only the IrisCode is stored.

When questioned about the wisdom of creating a system to accomplish everything from voting to welfare to everyday purchases, irrevocably tied to an individual person, all using public blockchains, Worldcoin argues “zero knowledge proofs!”

End of argument. Concerns assuaged. And indeed, some Worldcoin boosters seem satisfied with these very superficial answers, likely dazzled by the technical details that Worldcoin throws around with their posts about Gabor wavelets and phase-quadrant demodulation and poseidon hashes.

Excerpt from a Worldcoin post

But simply saying “we transform the iris data into something else” and “we’ll use zero knowledge proofs” should not be sufficient.

It is necessary to understand what kind of data can be leaked from the iris hashing algorithm — an algorithm that the team acknowledges is frequently changing. It is also necessary to understand what kinds of attacks could be enabled both on the network or on an individual participant if a malicious actor was able to obtain access to a participant’s data, ranging from a person’s WorldID account, to their IrisCode, to their full unhashed iris scan. This is a really critical issue, because there is no “password reset” when it comes to iris data. Questions about account recovery also remain unanswered — in previous reports, a person who uninstalled their World App was never able to regain access to their account.

There are also unanswered questions about the scalability of such an algorithm to the populations that Worldcoin is hoping to reach — past iterations of the algorithm have allowed one person to register multiple times, or have denied people access when they hadn’t already created an ID. While Worldcoin may have developed an algorithm that reliably distinguishes unique irises among its test pool, it’s not clear that it will work with a pool orders of magnitude larger.

As for zero knowledge proofs, Worldcoin trots this out as an answer to concerns about potentially connecting substantial amounts of sensitive data (transaction history and so forth) to a single permanent identifier. ZK proofs are a way of proving that something is true (e.g., “I have a valid World ID”, or “I’ve not yet received my WLD distribution this week”) without revealing additional details (e.g., which World ID is mine). But the implementation of this would be critical, and much of the burden here would lie on third parties — corporations, governments, etc. — that Worldcoin envisions adopting their system. A permanent record of potentially incredibly sensitive transaction histories, irrevocably linked to a biometric identifier, is a nightmare scenario. Worldcoin acknowledges this issue with no further elaboration: “While the Protocol is made to be used in a privacy-preserving manner, privacy cannot be enforced outside of the Protocol.”

(via @cryptograffiti)

Furthermore, Worldcoin relies heavily on each user’s ability and technical know-how to keep track of their private key, writing, “private keys need to remain private, as otherwise, a user can deanonymize themselves, even to actions they have performed in the past“.

Anti-surveillance advocate and Nym co-founder Harry Halpin describes Worldcoin’s knee-jerk “ZK proofs!” dismissal of privacy concerns — a frequent habit in the wider crypto world — as “zero-knowledge washing“: “taking a fundamentally evil concept or dubious concept and trying to make it look socially acceptable by adding some zero-knowledge fairy dust on top of it.” He also expresses doubts about the longevity of ZK proofs, expecting “quantum computing [to] break zero-knowledge proofs“ in five to ten years,

13though he and I differ somewhat on that particular prediction.

Those who resist Worldcoin’s initial rebuttals and continue to push the company on privacy concerns are then faced with Worldcoin’s next tactic: arguing that we already give up privacy in today’s society, so what’s a little more? Spokespeople trot out whataboutism with Apple and Google biometric scanning, and various backers argue that iris images are already widely distributed: “You have a headshot on your website. You walk around with your eyes open in front of cameras all day long.“

And if that doesn’t work? Well, there’s always investor Kyle Samani’s argument: “When it comes to Worldcoin, you don’t have to scan your eyeball. Like if you don’t want to, then fucking don’t.”

15This option is reasonable — indeed, quite tempting — if Worldcoin is relegated to the realm of the trivial, enabling NFT airdrops and the like. When it comes to more important use cases, “optional” biometrics suddenly become much more problematic. For example, in India, HIV patients have found themselves needing to submit their “optional” biometrics-linked Aadhaar identification number in order to access antiretroviral therapies.16

Tokenomics

It’s not yet clear how Worldcoin envisions WLD functioning. They refer to it as a “digital currency” which “could… become the widest distributed digital asset”. If Worldcoin is to function as a currency, as you might expect of an asset that’s being distributed in a universal basic income program, it would need to overcome the same types of issues that keep Bitcoin from functioning anything like a currency.

If it is to be more of a speculative asset that people hoard in hopes of the price going up, it seems ill-suited to Worldcoin’s UBI ambitions.

Furthermore, the initial token distribution looks a lot more like what you would expect out of the venture capital world than out of a public good organization. Worldcoin has generously reserved for insiders 25% of the WLD supply (up from an initial 20%, because development was evidently more “complex and costly” than anticipated).

It stands to reason that Worldcoin’s token distribution looks VC backed, and that’s because it is. In May, Worldcoin raised $115 million in a Series C round led by Blockchain Capital and joined by Bain Capital Crypto, Distributed Global, and, of course, Andreessen Horowitz.

18One wonders how they will balance their do-gooder mission with their need to generate massive returns for their backers.

Accessibility

Worldcoin’s loftier goals include “enabl[ing] global democratic processes”, providing global access to financial services, and even paying out AI-funded global universal basic income.

That is, if you have a smartphone and the technical know-how to use it, Internet connectivity, and access to an Orb. For Worldcoin’s more financial ambitions, people would also need access to an exchange where they could swap WLD for their local currency, or WLD would need to be widely adopted as a form of payment by merchants.

Today, an estimated 66% of the world uses a smartphone.

19Around the same percentage has access to the Internet, but this varies immensely by region, and is impacted by other factors including wealth and gender.20

Access to Orbs is a much more existential issue for the project: there are 346 Orbs out there in the world right now: that is one per every 23 million people. Worldcoin announced they will be rolling out

21more Orbs to reach a total number of 1,500 — meaning that then a mere 5.3 million people would have to travel to and line up per Orb.22Sam Altman has recently boasted (without evidence) that one person is being signed up every eight seconds.23With their claimed two million signups as a head start, at that rate they’ll have all 8 billion people in the world signed up by 2055 (assuming no population change, or change in rate of signup).

Finally, if Worldcoin truly wishes to onboard every person in the world, or be used for critical tasks, they will at some point have to grapple with the fact that not everyone has irises that can be scanned, due to factors including birth defects, surgeries, or disease.

Illegitimate access

“Show me the incentive and I'll show you the outcome,” says Charlie Munger.

What will happen when you promise people anywhere from $10 to $100 for scanning their eyeball? What if that’s not dollars, but denominated in a crypto token, making it appealing to speculators? And what if some people don’t have the option to scan their own eyeballs to achieve access to it?

A black market for Worldcoin accounts has already emerged in Cambodia, Nigeria, and elsewhere, where people are being paid to sign up for a World ID and then transfer ownership to buyers elsewhere — many of whom are in China, where Worldcoin is restricted. There is no ongoing verification process to ensure that a World ID continues to belong to the person who signed up for it, and no way for the eyeball-haver to recover an account that is under another person’s control. Worldcoin acknowledges that they have no clue how to resolve the issue: “Innovative ideas in mechanism design and the attribution of social relationships will be necessary.“ The lack of ongoing verification also means that there is no mechanism by which people can be removed from the program once they pass away, but perhaps Worldcoin will add survivors’ benefits to its list of use cases and call that a feature.

Relatively speaking, scanning your iris and selling the account is fairly benign. But depending on the popularity of Worldcoin, the eventual price of WLD, and the types of things a World ID can be used to accomplish, the incentives to gain access to others’ accounts could become severe. Coercion at the individual or state level is absolutely within the realm of possibility, and could become dangerous.

Conclusion

Worldcoin seems to be embracing a modified version of the “move fast and break things” mantra that has become all too popular in the tech world. “Build a massive database of biometric data and then figure out what to do with it someday” is a little less catchy, though.

The issues with Worldcoin that I list here are far from exhaustive, and I’ve included some further reading below from others who’ve shared their thoughts.