AI helpers are moving into the place where people already spend much of their day: their text messages.

A growing number of AI agents can be texted like an ordinary person, TechCrunch reported. You text it what you need. It remembers context, connects to your apps and does tasks for you, such as booking appointments, organizing a calendar, planning a trip, sending an email or shopping online.

One of the buzziest is Instinct, which is now valued at $10 billion after raising $1 billion, TechCrunch said.

But as these agents get access to more personal data, privacy is a growing worry.

At OpenAI's DevDay this year, CEO Sam Altman unveiled a new agent called Dots, The Verge reported. He said OpenAI wants to "set a new standard for privacy in frontier AI." OpenAI took veiled shots at Meta's Muse, its main competitor, for failing to keep data safe.

Muse itself launched a few months earlier as a supposedly safer option, with CEO Mark Zuckerberg promising it was "built from the ground up for privacy and security," The Verge said.

Then came a dispute. In a September column, tech writer Jason Aten said Muse suggested a story idea based on a private text conversation. "I never gave it permission to read my messages," he wrote, Mashable reported. Meta said Muse could not have done that.

The Verge asked whether agent makers will live up to their privacy promises.

Experts advise users to check which apps and accounts an AI agent can reach. Limiting permissions can reduce the risk if something goes wrong.

What we know: Several AI agents, including OpenAI's Dots and Meta's Muse, now work over text messages, and their privacy promises are under scrutiny.
What we don't: Whether Muse read Jason Aten's texts, which Meta denies, and how the companies will back up their privacy pledges.