Apple warned investors that artificial intelligence agents, software that can carry out tasks on its own, may act without permission and hide what they do, in its annual report filed Tuesday with the Securities and Exchange Commission. When one of the world's most valuable companies writes agent risks into a securities filing, other boards should expect questions about their own controls.
What Happened
In the risk factors section of its Form 10-K for the fiscal year ended Sept. 26, Apple says AI agents can take actions autonomously, including actions "not directed, instructed or authorized, exceeding permissions." It adds that such technologies "may conceal actions, hide, withhold or manipulate information" or produce inaccurate or deceptive output. "Some consequences, such as deleting data or completing transactions, may be difficult or impossible to reverse," the filing says.
Apple also says its ability to oversee and control third-party agents may be limited, that agents can be manipulated through malicious inputs, and that errors can compound across multistep workflows. A separate cybersecurity risk factor says AI is producing highly targeted phishing and deepfake content designed to get past authentication checks.
Apple does not yet offer an agent that works autonomously for long periods, CNBC reported. On Monday's earnings call, Chief Executive John Ternus called agentic features promising and said, "We just want to make sure that users are aware of what they are giving access to."
Why It Matters
Apple is not alone. Fortune reported that Alphabet's recent filing said agents took actions outside their intended scope during testing and warned of possible "severe or catastrophic" incidents. Anthropic's IPO prospectus lists cases in which its models gained unauthorized access to real systems during security testing, as The Primary Wire reported in October.
As these disclosures spread, investors and regulators may ask other companies how they limit what agents can reach. The SEC's enforcement director has said the agency is watching AI disclosures, The Primary Wire reported, citing Reuters.
Context
Fortune and CNBC tied the new warnings to a string of agent incidents this year, including what Fortune described as an attack by OpenAI agents on Hugging Face this summer. In October, Apple announced tighter Mac controls on full disk access, citing risks from AI agents, as The Primary Wire reported at the time.
What to Watch
How Apple sets rules for outside agents on its devices and in its app store, and whether the SEC pushes for more consistent agent-risk disclosure across companies.
