Apple warns that AI agents may act in unexpected or unpredictable ways
Apple warned in the filing that AI agents can autonomously interact with internal or external systems, execute or modify code, send communications and complete transactions, and may act in unexpected or unpredictable ways, including taking actions that were not instructed or authorized. Apple said the company's ability to oversee and control third-party agents may be limited, and it adjusted Mac's Full Disk Access permissions last month to reduce the related risks. New CEO John Ternus said on the earnings call that agent capabilities are a promising tool for users and that the company has room to expand Siri, but it needs to ensure users clearly understand what access they are granting.
Why read this
AI score 58
Apple flagged in a regulatory filing the risks of agents' unauthorized access and unpredictable behavior, and has adjusted Mac disk permissions; this can be viewed alongside its Siri agent plans to understand the security boundaries.
Topics
Body · AI translation
“In particular, AI agents and other technologies can take actions autonomously, including interacting with internal or external systems—which in some cases they may not be authorized to access—executing or modifying code, sending communications, and completing transactions,” Apple wrote in the filing.
The company added that agents “may act in unexpected or unpredictable ways, including taking actions that were not instructed, directed, or authorized.”
Earlier this year, OpenAI released Dots and Meta launched Muse. Both AI agents use advanced AI models to complete tasks autonomously. They can even email users or alert them when new messages arrive.
Apple made this disclosure after several incidents in which runaway AI agents broke out of sandbox environments and attacked third-party websites, such as the Hugging Face incident, which has renewed attention among policymakers in Washington, D.C., and around the world.
In addition, agents’ ability to use computers the way humans do is forcing large companies such as Amazon and JPMorgan Chase to set policies around agent access to services for strategic or security reasons.
Last month, Apple announced changes to “Full Disk Access” on the Mac to reduce the “risks associated with such access” for AI agents installed on PCs.
“The company’s ability to oversee and control third-party AI agents and other technologies may be limited,” Apple warned in the filing.
Apple currently has no agentic AI. But when new CEO John Ternus was asked about AI agents on Monday’s earnings call, he said the company has “real room” to expand Siri’s capabilities.
“In terms of agentic capabilities, we think these are promising tools for users,” Ternus said. “We just want to make sure users are clear about what access they are granting. We want to make sure we do this in a way that protects security and privacy.”