AI agent makers are promising privacy — will they deliver?

At OpenAI's DevDay, CEO Sam Altman unveiled the new AI agent Dots, claiming OpenAI wants to "set a new standard for privacy in frontier AI," while taking veiled shots at Meta's competing agent Muse for failing to keep user data safe. Yet Meta had made similar promises when launching Muse as a safer alternative to OpenClaw, with Mark Zuckerberg saying it was "built from the ground up for privacy and security." Despite Muse's rapid growth to 600,000 daily active users in the US, its promises appear to have fallen short: Meta can still access user data stored on a secure VM, a zero-day vulnerability was quickly exposed, and serious security issues reportedly emerged before launch, including one that could have let users access Meta's internal databases. Muse also liberally collects data, defaulting to allowing Meta to train models on user inputs, and it created uncomfortable situations such as uploading private messages or sharing a user's address. OpenAI has seized on these failures, positioning Dots as more trustworthy with user controls and zero data retention options for enterprises, but because Dots is limited to expensive subscription tiers, fewer people use it and fewer privacy scandals have surfaced. The article notes that AI agents inherently require vast amounts of personal data, making some users uncomfortable, and that not all companies—like Instinct, criticized for overly broad terms of service—are making privacy promises. Ultimately, AI labs appear to be betting on a three-part strategy: make agents useful, make them cute and disarming, and make privacy promises—hoping they hold up.

(Source：The Verge)