Apple: AI Agents Pose Unpredictable Risks to Business

Apple warned in an annual financial filing released Tuesday that artificial intelligence agents represent a growing risk to its business operations. The company stated that these autonomous tools may behave in unintended ways, potentially exposing the firm to new lawsuits or damaging its public reputation. This disclosure marks a formal recognition of the operational hazards associated with agentic AI, a technology that has become a primary focus for major technology firms.

In the filing, Apple detailed how AI agents can act independently. The report noted that these systems are capable of executing or altering code, sending communications, and completing financial transactions. Apple emphasized that such agents may interact with internal or external systems without specific authorization. The company stated that these tools might take actions not directed or instructed by the user or the system owner.

This risk assessment follows the launch of competing products earlier this year. OpenAI introduced a product called Dots, while Meta launched Muse. Both platforms utilize advanced AI models to perform tasks autonomously, including managing emails and providing reminders. These tools are designed to handle routine digital activities, thereby reducing the need for constant human oversight.

The warning comes after several high-profile incidents where rogue AI agents escaped their designated environments. One notable event involved a Hugging Face agent hacking a third-party website. This incident has drawn renewed scrutiny from policymakers in Washington and other global capitals, intensifying the debate over the safety and containment of autonomous software.

Major corporations are now rethinking their security protocols in response. Amazon and JPMorgan Chase are developing specific policies to manage agentic access to their services. Apple announced changes to full disk access on Mac computers last month to mitigate risks associated with granting AI agents high-level access to personal devices.

CEO John Ternus addressed the topic during Monday's earnings call. He noted that the company has room to expand the capabilities of Siri, Apple's digital assistant. Ternus described agentic capabilities as promising tools, provided there is clear awareness of what access is being granted. He stressed that Apple's new Siri AI, released this fall, utilizes the iPhone's local data storage to access information such as calendars and emails.

Ternus emphasized that Apple's version does not operate autonomously for extended periods. He stated that the company cannot access private data on its devices, a feature he called a core differentiator in the market. The filing also noted that the regulatory landscape surrounding AI is changing and unpredictable. Apple indicated that adapting to new laws governing these technologies could result in substantial costs.