From The Verge · Hayden Field · · 2m
AI agents promise privacy. Nobody knows if they'll keep it.
OpenAI and Meta are locked in a race to convince users their AI agents are safer than rivals', but their track records suggest otherwise.
In brief
OpenAI and Meta are locked in a race to convince users their AI agents are safer than rivals', but their track records suggest otherwise. OpenAI and Meta are competing by promising privacy for their AI agents, but Meta's Muse has already suffered security breaches and data collection issues that undermine those…
OpenAI launches Dots with privacy pitch
At OpenAI DevDay on October tenth, CEO Sam Altman unveiled Dots, the company's new AI agent, and promised to set a new standard for privacy in frontier AI.
“set a new standard for privacy in frontier AI”
Meta's Muse promised the same thing
Meta launched Muse a couple of months earlier as a supposedly safer alternative to its predecessor OpenClaw. CEO Mark Zuckerberg promised it was built from the ground up for privacy and security.
“built from the ground up for privacy and security”
The privacy arms race begins
AI labs are trying to convince users to share more personal data with their agents by promising better privacy than competitors. The question is whether companies can actually keep these promises.
How Meta said it would protect data
Meta said user data is stored on a secure virtual machine, an isolated Linux computer with a browser, CPU, memory and storage.
Muse's early success masked problems
Muse topped the App Store charts and gained six hundred thousand daily active users in the US within weeks. But Meta's privacy promises quickly fell short as security issues emerged.
Security breaches exposed Muse's flaws
A security researcher quickly exposed a zero-day vulnerability that could allow someone to take control of Muse.
Muse collects and shares user data
Muse defaults to allowing Meta to train models on what users put into it.An Inc.
OpenAI claims it learned from Meta's mistakes
OpenAI's Glen Coates said the company is in a different position to Meta because they don't have an AI product with one point two billion users.
“they don't have an AI product that has 1.2 billion users”
Dots offers enterprise privacy controls
OpenAI presented a framework allowing enterprises to have stronger controls over their data and zero data retention policy options, meaning no data is stored on OpenAI servers.
Users remain wary despite promises
AI agents require so much personal data that users still feel uncomfortable sharing sensitive information.
The takeaway
OpenAI and Meta are competing by promising privacy for their AI agents, but Meta's Muse has already suffered security breaches and data collection issues that undermine those claims.
