The most dangerous attendee on a bank’s next video call won’t be on the meeting invite. Deepfake fraud has gone from conference-room legend to a budget line: Deloitte’s Center for Financial Services projects AI-enabled fraud losses in the US could reach $40 billion by 2027, up from roughly $12.3 billion in 2023. That is the number Sam Altman’s identity project is betting banks will pay to shrink.
On Sunday, World — the identity network Altman co-founded — launched World ID for banking and finance at Money20/20 in Las Vegas, the first time it has aimed its verification stack at a single industry. Co-founder Alex Blania unveiled a bundle of three existing products repackaged for financial firms, The Block reports: the World ID app, an identity wallet for storing and presenting what World calls “Proofs”; World ID for Zoom, which checks that the person on a video call is a real human and not a deepfake; and World ID for Agents — the tool launched as AgentKit — which lets a verified person hand those credentials to an AI agent. The target: customer accounts, employee workflows, and the AI agents now operating inside both.
The Piece That Actually Ships
The genuinely new part is the Zoom check. World ID for Zoom is now generally available after a limited beta that began in April, and Tools for Humanity — the primary developer behind World — has joined Zoom’s independent software vendor exchange program. Unchained reports the check compares the image captured when a user first verified at World’s Orb device with a live selfie and the meeting video — and any participant can request one mid-call. VanEck was the only finance firm World named among the April beta testers. World says it will demo the technology at Zoomtopia, Zoom’s conference opening October 20, and at Cloudflare Connect.
The Catch: No Banks, No Prices, No Timeline
Here is what the launch doesn’t include: a bank. World named four design partners — OpenExchange, Bain Capital, VanEck Funds, and Outtake — that will “explore” how the technology could serve financial services. No bank is using the product, no pricing was disclosed, and no rollout timeline exists. That makes Sunday’s news a product announcement aimed at banking, not a banking product with customers. And the “more than 18 million people” holding a Proof of Human are consumers who verified at Orbs worldwide — not bank-vetted users. “As AI becomes increasingly prevalent in all aspects of our lives, enterprises are requesting individuals provide ‘Proofs’ about themselves to help mitigate fraud and increase trust,” Tools for Humanity President Sandro Herbig said in the announcement.
Why this matters
Look past the fraud framing and there is a stranger bet. By packaging AgentKit for banks, World is normalizing a future in which verified humans delegate their proof-of-personhood to AI agents — which then present it to third-party platforms. The same system promising to keep bots off your Zoom call is building the infrastructure for bots to transact with your credentials. That is not necessarily a contradiction — delegation with limits is how corporate banking already works — but it puts the trust question exactly where banks are least comfortable: who controls the handoff, and what happens when the agent holding your proof gets compromised. World wants banks to buy the answer before they’ve asked the question.
Meanwhile in Apps & Tools: Google is letting you keep your vibe-coded apps private on the Play Store, and OpenAI’s Codex cloud explained.
