Pro Security Tech Radar

Agentic AI has “profound” issues with security and privacy, Signal President says

March 10, 2025
admin

Click here to visit Original posting

  • Signal President Meredith Whittaker calls out dangers of AI agents
  • Increased work task usage means "you’re not doing any of that yourself"
  • The tools have already raised privacy and security concerns

The president of popular messaging app Signal has warned AI agents come with a significant risk to privacy and security, which is “haunting” the hype around agentic AI.

Speaking at SXSW, Meredith Whittaker argued AI agents are being marketed like a "magic genie bot” which thinks many steps ahead and completes tasks for users, so, “your brain can sit in a jar, and you’re not doing any of that yourself”.

But this comes at a cost, and although the use of AI agents is proving popular, Whittaker emphasized that there’s a “real danger” with these bots, because they require extensive access to user data.

Excessive access

Say you asked an AI agent to book a concert for you and your friends - this should be a fairly straightforward task, but it does mean the AI agent would need access to your browser, your credit card information, your calendar, and even your messages to let the friends know.

This all means that with just one task, the agent now has access to your financial details, your day to day plans, and your messages - which could be extremely damaging if the data fell into the wrong hands.

"It would need to be able to drive that across our entire system with something that looks like root permission, accessing every single one of those databases, probably in the clear because there's no model to do that encrypted," Whittaker explained.

Messaging apps like Signal which have end-to-end encryption (E2EE) would compromise the privacy of user messages if integrated with AI agents, even if this was just to text friends or summarize your incoming messages, she noted.

“That’s almost certainly being sent to a cloud server where it’s being processed and sent back,” Whittaker adds. “So there’s a profound issue with security and privacy that is haunting this hype around agents, and that is ultimately threatening to break the blood-brain barrier between the application layer and the OS layer by conjoining all of these separate services [and] muddying their data,”

Via TechCrunch

You might also like