The Apple Watch Series 12 and the Apple Watch Ultra 4 were presented this week. Although external changes are minimal, the devices bring notable functionalities, including the Siri Recap feature, which listens to dialogues in the user's environment to create summaries.
In light of concerns regarding the security of a watch with active listening capabilities, John Ternus, who assumed the role of Apple's executive president at the beginning of the month, offered assurances. In interviews given to TechRadar and Tom's Guide, the executive emphasized that the company maintains a focus on security and privacy in all its actions.
Among the innovations of the Apple Watch Series 12 and the Apple Watch Ultra 4 is Audio Intelligence, a set of tools that includes the Live Rewind function. The latter allows retrieving the last fifteen seconds of nearby speech and generating its corresponding transcription.
Siri Recap works by allowing the smartwatches to monitor ambient conversations so that, assisted by artificial intelligence, they can produce notes or summaries of these interactions. However, the question arises about the comfort of the user or third parties knowing that the watch is recording.
In response, John Ternus and Greg Joswiak, Apple's Senior Vice President of Marketing, clarified that Siri Recap does not perform complete recording or transcription of conversations. For this reason, the content cannot be accessed by hackers or shared with third parties.
Additionally, the technology does not perform speaker attribution, meaning it does not directly identify who spoke or which people are nearby in the generated summaries, preventing the watches from pointing out who made a specific statement.
To reinforce protection, Siri Recap filters speeches that might be sensitive or harmful, such as those containing hate speech, financial information, or personal identification data, such as document numbers.
Another concern raised relates to the handling of conversation data through Private Cloud Compute, Apple's cloud infrastructure used to process AI tasks. The company assures that this solution was developed to provide a level of security and privacy equivalent to processing done locally.
Apple further details that the initial processing of conversations occurs directly on the Apple Watch, using the Secure Enclave, a security compartment linked to the device's chip, specifically the new S11. The audio processed in this location is immediately discarded, and only a synthesized text is sent to Private Cloud Compute for summary generation.
It is important to note that Apple will continue to defend its practices. Therefore, waiting for independent tests is crucial to confirm whether the Audio Intelligence features are effectively functional and secure. However, more cautious users have the option to disable these new functionalities.
