---
title: "Apple Warns Investors in Annual Report: AI Agents Could Act Unpredictably and Harm Business"
description: "In its annual financial report, Apple included AI agents in its list of corporate risks, pointing to their potential to act unpredictably. The company has already altered Full Disk Access policies on Macs, and CEO John Ternus spoke about Siri's \"significant potential\" in the agentic space."
date: 2026-11-04T23:31:29.197Z
lang: en
url: https://xab.info/en/posts/apple-ai-agents-risks-financial-report
tags: [apple, ai-agents, siri, financial-report, data-privacy, macos, full-disk-access]
author: "Alexander Buran"
publisher: "XAB.info"
---

# Apple Warns Investors in Annual Report: AI Agents Could Act Unpredictably and Harm Business

![Blue Apple iPhone smartphone with triple camera module resting on a wooden surface](https://xab.info/media/2026/11/04/apple-ii-agenti-riski-finansovy-otchet/apple-ii-agenti-riski-finansovy-otchet-1.webp)

## Key Points

- Apple's annual financial report lists AI agents as a corporate risk capable of leading to lawsuits and reputational damage
- In October 2026, Apple modified the Full Disk Access mechanism on Macs to mitigate AI agent-related risks
- CEO John Ternus called agentic features "promising tools" for Siri while emphasizing that user data remains on the device
- Apple pointed to the unpredictability of AI regulation and the risks of negative consumer perception regarding automation

In its annual financial report submitted to investors, Apple has for the first time explicitly listed among corporate risks the potential actions of autonomous AI agents capable of operating in an "unforeseen or unpredictable manner." According to company leadership, such scenarios pose a direct threat to the business, ranging from lawsuits to severe reputational damage. The document emphasizes that AI agents "may take actions autonomously: interact with internal or external systems they may not have permission to access, execute or modify software code, send messages, and conduct transactions." Apple believes this combination of capabilities turns agentic systems into a potentially destructive factor for any organization implementing such technologies.

### Proactive Response: Changes to Full Disk Access on Mac

To mitigate risks associated with the use of AI agents, Apple introduced changes in October 2026 to the Full Disk Access mechanism on Mac computers. This decision was a practical step addressing the threats the company itself outlined in its financial report. Essentially, Apple acknowledges that even though it does not yet have its own agentic AI systems, the ecosystem around its devices is already encountering third-party programs capable of requesting broad privileges. Altering the disk access policy is an attempt to limit the attack surface and give users more transparent control over which applications access sensitive data.

### Siri as a "Promising Tool": John Ternus's Stance

Although Apple does not yet possess its own agentic AI systems, CEO John Ternus publicly stated that the company has "significant potential" to expand the capabilities of the Siri voice assistant. He described agentic features as "promising tools for users." At the same time, Ternus emphasized a key principle: "We simply want to make sure users understand what exactly they are granting access to. Our job is to guarantee that all of this happens in compliance with security and privacy requirements." The commercial logic is also clear: Apple expects that new AI features in Siri will serve as an additional incentive for device upgrades and help offset growing expenses associated with the global AI boom.

### Fundamental Difference: Data Stays on the Device

Ternus specifically highlighted that Apple's approach fundamentally differs from competitors' strategies. "Your iPhone stores a huge amount of personal data — contacts, correspondence, health information, financial details, and much more. Siri can access this data, helping you work with it as securely and privately as possible — in such a way that no one else gets access to it," he stated. In contrast, commercial AI agents from companies like OpenAI or Meta request broad access to user accounts, including email, to enhance efficiency. Thus, Apple positions its model as "privacy by default," whereas competitors build agentic services based on centralized access to user data.

### Regulatory Uncertainty and Ethical Risks

In addition to technological threats, Apple noted in the report that regulation in the field of artificial intelligence and "superintelligence" is actively changing and remains unpredictable. Adapting to new laws concerning AI could require significant additional costs for the company. The document separately outlines risks associated with ethical concerns and negative consumer perceptions of automation, artificial intelligence, and superintelligence — including fears regarding security threats and other societal issues. For a company whose brand is built on trust and privacy, such market sentiment poses a threat no less serious than technical failures of agentic systems.

Taken together, Apple's 2026 financial report marks a turning point: the major tech giant openly acknowledges that autonomous AI agents are not only a growth tool but also a source of systemic risks requiring both engineering and regulatory measures. For investors, this is a signal that the company is factoring expenses for security, legal protection, and adaptation to emerging legislative changes into its financial models.

## Fact-Check & Sources

- [Apple предупредила инвесторов: ИИ-агенты могут выйти из-под контроля и навредить бизнесу window-new](https://3dnews.ru/1150861/apple-predupredila-investorov-iiagenti-mogut-viyti-izpod-kontrolya-i-navredit-biznesu) - The primary source contains the full text of Apple's financial report summary, featuring quotes from John Ternus, descriptions of Full Disk Access changes, and regulatory risk analysis. All facts in the article correspond to this source.

## FAQ

### What specific AI agent risks did Apple outline in its financial report?

Apple noted that AI agents can act in an "unforeseen or unpredictable manner," interact with systems without access rights, modify software code, send messages, and conduct transactions. This creates the risk of lawsuits and reputational damage.

### Does Apple have its own agentic AI systems?

No, at the time the report was published, Apple does not have its own agentic AI systems. However, CEO John Ternus stated there is "significant potential" to expand Siri in the agentic direction.

### What did Apple do to mitigate risks associated with AI agents?

In October 2026, Apple introduced changes to the Full Disk Access system on Mac computers, narrowing the surface of potential threats.

### How does Apple's approach to AI differ from OpenAI and Meta?

Apple emphasizes that it does not access users' confidential data on their devices, and Siri processes data locally. Conversely, commercial AI agents from OpenAI and Meta request broad access to user accounts, including email.