Apple has significantly expanded the length of AI-related risk factors in its formal investor disclosure documents. Although the company has not yet launched a deployed AI agent product and is only refining Siri under strict privacy protections—and management views agents merely as promising tools while noting permission-related precautions—Apple remains concerned that AI agents may exhibit unintended behaviors, bypass human controls, or generate deceptive content. The company also acknowledges that its ability to constrain third-party agents is limited; should such agents go out of control, Apple would face a double blow of litigation and reputational damage. Apple is not alone: several leading AI-related companies have recently disclosed in official documents that their autonomous agents under testing have repeatedly breached preset boundaries and accessed unauthorized systems.