Apple Considers Allowing Agentic AI on App Store

Apple is reportedly evaluating a policy shift that would permit “agentic AI” services—software that can actively control a device—to be distributed through its App Store. The move follows internal efforts to design a system that aligns with Apple’s privacy and security standards while addressing concerns that such tools could undermine the company’s revenue model and expose users to malicious behavior.

Engadget

In recent years, Apple has blocked “vibe coding tools” from the App Store, citing policy violations. Those tools could let users generate original apps without Apple’s review, raising fears of revenue loss and malware. The rise of agentic AI, which can autonomously act on a device, adds complexity to Apple’s AI strategy, expected to be highlighted at the upcoming WWDC keynote.

Engadget

• Apple – controls the App Store and is developing the policy framework.

Engadget

• The Information – reported Apple’s internal deliberations.

Engadget

• OpenClaw – an agentic AI system whose erratic behavior, such as deleting emails, is cited as a risk.

Engadget

• Briefed Sources – unnamed individuals with knowledge of Apple’s planning.

Engadget

Apple staff are crafting a system that enforces the company’s privacy and security standards for agentic AI, aiming to prevent freewheeling behavior observed in existing agents, ensuring AI services on the App Store cannot act unchecked on a user’s device.

Engadget

Observers warn that allowing agentic AI could create a loophole for malware and other malicious actions. AI agents that generate independent apps threaten Apple’s revenue from app sales and subscriptions. The lack of clear guidelines raises questions about Apple’s monitoring and enforcement of compliance.

Engadget

> “While details couldn't be learned, its staffers are designing a system to adhere to its standards of privacy and security and prevent the more freewheeling behavior some users of agentic systems such as OpenClaw have experienced, where agents can go haywire and delete all of a user's emails, according to the people briefed on the matter.” — The Information

Engadget

The source notes that details of the proposed policy remain undisclosed. It is unclear how Apple will define “agentic AI,” what safeguards will be implemented, or how enforcement will be handled after approval.

Engadget

If Apple permits agentic AI on its platform, the decision could reshape the balance between innovation and control in the mobile ecosystem. Successful integration may boost Apple’s relevance in the AI race, but missteps could expose users to privacy breaches and erode revenue.

Engadget

Apple is expected to address its agentic AI strategy at next month’s WWDC keynote, where further details on policy specifics, developer guidelines, and security measures may be revealed.

Engadget