A recently addressed vulnerability in Apple’s audio service permitted applications on iPhones and iPads to record audio without user consent, even after permission was explicitly denied through the microphone settings. The initial security patches addressing this issue were released on August 17. Both Apple and the security researchers involved have confirmed that there is no evidence of this flaw being exploited in real-world scenarios.
Research and Discovery
The technical analysis of the flaw was shared by Nik Tsytsarkin from Meta Red Team X, with a parallel examination conducted by Mustafa Calap, a security engineer, released on the same day. Notably, an orange indicator, designed to show when the microphone is in use, failed to activate during tests on specific app builds. This issue persisted even when users declined microphone access on initial prompts.
Tsytsarkin’s investigations revealed that the system logs did not register any microphone usage by the app, as it effectively bypassed the intended controls. Calap’s tests further showed that the unauthorized recordings occurred without the typical indicators appearing in the iOS interface, such as the Dynamic Island or Control Center.
Technical Implications and Updates
The security flaw was documented as CVE-2026-65339 and affected a range of devices. The updates to address the vulnerability were released across multiple Apple systems, including iOS and iPadOS 26.6.1, macOS Tahoe 26.6.2, and others, between August and October. Importantly, the flaw did not extend additional privacy permissions beyond audio capture, and no settings existed to block the vulnerability aside from applying the updates.
Both researchers indicated that the apps used in their tests were sandboxed and lacked additional entitlements, relying on the NSMicrophoneUsageDescription privacy key for microphone prompts. The recordings could commence when the app was active and continued even when the device was locked.
Security Measures and Future Outlook
To prevent similar issues, Apple has implemented a fix requiring verification of the caller’s identity within its audio service. This ensures that only authorized processes can initiate recordings, closing the loophole that previously allowed apps to impersonate the audio service itself.
As a precautionary measure, users are advised to keep their devices updated with the latest security patches. While the flaw has been assigned a CVSS score of 5.0 by CISA, it remains absent from the agency’s list of known exploited vulnerabilities. Apple’s transparency notes emphasize the importance of these updates to safeguard sensitive user information against unauthorized access.
In conclusion, while the vulnerability posed a significant risk to user privacy, the swift response and corrective measures undertaken by Apple and the researchers highlight the ongoing commitment to security within the tech industry.
