Dots and Muse: Are OpenAI and Meta Keeping Their Privacy Promises?

TVS News AI, editorially supervised, presents Dots and Muse: Are OpenAI and Meta Keeping Their Privacy Promises.

Sam Altman took the stage at this year's OpenAI DevDay and presented the new AI agent Dots. But instead of just showcasing features, he played the privacy card: OpenAI wanted to "set a new standard for privacy in frontier AI." The company spent the rest of the day repeatedly glancing at rival Meta — over what it claimed was weak data protection at the competitor.

Their agent Muse had only launched a few months earlier — as a supposedly secure alternative to its predecessor OpenClaw. CEO Mark Zuckerberg had sold the product back then as being privacy-friendly from the ground up. Now Meta finds itself in the crosshairs of a competitor that is itself touting big privacy words.

When the world's most powerful AI labs court your personal data, a lot of it sounds like privacy-first. The Verge asked: What really lies behind the promises? We break down what you need to know — and what to watch out for before entrusting your data to Dots or Muse.

What exactly have Dots and Muse promised?

Altman was clear on stage. OpenAI wanted to treat data protection not as a feature, but as a foundational principle. That's a claim that deserves attention — because AI agents dig deeper into your life than classic chatbots do. They take on tasks, read calendars, search through emails. Anyone using such tools inevitably reveals more than most people are comfortable with.

set a new standard for privacy in frontier AI

Meta countered with Muse. Zuckerberg positioned the product as a direct response to growing concerns among many users that their data could fall into the wrong hands at AI services. The message: With us, you're safer than with OpenAI.

built from the ground up for privacy and security

Why the data protection dispute is heating up right now

AI agents are the next stage of digitalization. They act autonomously — and to do so, they need access to sensitive data. In an era where data leaks and cyberattacks are the order of the day, that's a delicate promise. If a provider cuts corners here, the damage can far exceed what a hacked email account causes.

What's notable at The Verge: both camps accuse each other of not taking users' data seriously enough. OpenAI criticized Meta for inadequate protection. Meta points to its supposedly secure architecture. In the end, users are caught in the crossfire — and left to believe whomever they want.

What this means for your data in concrete terms

Anyone using an AI agent like Dots or Muse typically grants the system far-reaching permissions. The agent sees appointments, contacts, possibly your inbox. That makes it an attractive target for attackers — and a privacy risk that goes beyond what conventional apps have collected so far.

What to watch out for before choosing either agent: What data is actually stored — and for how long? Can you delete it? Who has internal access? As long as OpenAI and Meta don't translate their promises into verifiable features, the phrase "we take data protection seriously" remains above all: a marketing promise.

What matters in the end: deliver, don't promise

The starting position is clear: OpenAI and Meta are in a race for the safest AI platform. Both tout big words. What's missing so far are independent audits that show whether Dots and Muse actually keep their promises. Until then: trust is good, control is better. Read the privacy settings carefully before giving an AI agent access to your digital life.