Apple has announced a initiative to bolster the confidence in the security of its upcoming AI system, Apple Intelligence. The tech giant is offering a bounty of up to $1 million to anyone who can successfully hack its new system, which is set to launch on October 28.
The announcement was made on Thursday, inviting “all security researchers—or anyone with interest and a technical curiosity“ to independently verify the security claims surrounding Apple Intelligence. The company is particularly focused on its Private Cloud Compute (PCC) system, which is designed to handle complex AI tasks that cannot be processed on-device. Apple claims that this system features end-to-end encryption and deletes user requests immediately after processing.
Apple’s bounty program includes various payout tiers for different types of discoveries. The highest reward of $1 million is reserved for anyone who can execute code on the system without detection and access sensitive areas. Other rewards include $250,000 for accessing user request data outside the designated trust boundary and $100,000 for executing unverified code. A smaller reward of $50,000 is available for accidental data disclosures. Apple went even further, saying that it might give a reward for a discovered vulnerability ‘even if it doesn’t match a published category’.
The company has previously allowed only select third-party auditors to examine its PCC system. However, this new initiative opens the door for a wider range of participants to test the system’s security. Apple is providing a security guide and access to a virtual research environment for those interested in participating. To access this environment, individuals will need a Mac with an M-series chip and at least 16 GB of RAM.
Apple Intelligence is set to debut with the iPhone 16, 15 Pro, and 15 Pro Max. The AI system will introduce several new features, including enhanced writing tools, smart replies, and a redesigned Siri. However, some features, such as ChatGPT integration and advanced image processing capabilities, will be rolled out in future updates.
The company has emphasized its commitment to user privacy, stating that the PCC system is designed to ensure that user data remains secure and private. Apple has a long-standing reputation for prioritizing privacy, and this initiative aims to reinforce that image as it expands its AI capabilities.
In a statement, Apple expressed confidence in its security architecture, stating, “We believe Private Cloud Compute is the most advanced security architecture ever deployed for cloud AI compute at scale.” The company stated that it looks forward to collaborating with the research community to enhance the security and privacy of its systems over time.
With the launch date fast approaching, Apple is preparing for potentially high demand for the new AI features. Users will be able to join a waitlist once the system goes live. The initial version of Apple Intelligence will be compatible only with US English, with support for other languages expected to follow in December.
(Photo by Medhat Dawoud on Unsplash)
