Apple has warned that autonomous AI agents could access systems without authorization, bypass security controls and carry out actions that are difficult to reverse.
The risks appear in the company’s 2026 Form 10-K filed with the U.S. Securities and Exchange Commission on November 3. The disclosure is a risk warning, not a report of a specific incident.
Apple says AI agents may interact with internal or external systems, execute or alter code, send communications and complete transactions. The filing warns that these tools, whether developed by Apple or third parties, may exceed permissions, resist instructions, circumvent security measures or be manipulated through malicious inputs. They may also conceal actions, withhold or manipulate information, or produce inaccurate and deceptive outputs.
The company says agent behavior can be difficult to anticipate, control, test and audit, while errors may compound across multistep workflows. It warns that consequences such as deleting data or completing transactions may be difficult or impossible to reverse. Apple also says its oversight of third-party agents may be limited and its policies may not keep pace with rapidly advancing tools. Apple has already added agentic coding capabilities to Xcode, allowing compatible agents to inspect projects, run builds and make changes.
If this article helped you, please consider supporting our work. Every small contribution keeps Abijita.com independent and running.
