Lawyers in four all-party-consent states say wearer could face criminal risk from Apple Watch audio features

Four people at a table, and one of them double-presses the Digital Crown to catch a phone number. Under the statutes in California, Massachusetts, Pennsylvania and Washington, the consent that matters belongs to the other three ^[5]. Joe Simons, a criminal defence attorney in Boston, said the nuances have not been litigated in Massachusetts, and he expects the issue to reach the state as a case of first impression ^[12]. Apple's Audio Intelligence overview sets out where the text comes from. Microphone audio runs into a buffer inside the Secure Exclave on the S11 chip, and new audio continuously overwrites the old. The double-press sends the preceding 15 seconds to the paired iPhone, on-device speech recognition turns it into text, and the phone discards the raw audio ^[17]. Because the overwrite never stops, the last 15 seconds of nearby speech sits in the buffer before anyone touches the crown, and the press exports it ^[2]. Apple says the features do not create an audio file ^[10]. The attorneys raised something narrower than recording. Speech processed continuously, with no conventional recording at the end of it, may still count as interception under state law ^[11]. Two of them read the same statutes in opposite directions. Phil Lauer, a defence attorney in Pennsylvania, said the technology is "very close to being legal, because it's not an individual who is actually receiving and using this data" ^[13]. He added that the opposite case can be made too, since the device picks up audio and interprets it, and said the courts will have to take the question up ^[14]. Mark Blair, a principal at the Seattle firm Blair Kim Moeller, said interception is not limited to the person who keeps a recording for later playback. A wearer could face a criminal charge after the content is gone from the watch, he said ^[15]. Blair said a wearer has to announce that they have an Apple Watch and that it is recording. "Absent that, it is a criminal violation," he said ^[16]. The four states Lung called are joined by around half a dozen more with the same all-party requirement, including Illinois and Florida ^[7]. Call it ten ^[1]. Apple has kept the audio small and local. A lightweight model on the S11 decides whether speech is happening nearby, and the company says that model does not transcribe or record ^[18]. Detected audio is encrypted, moved to the Secure Exclave on the paired iPhone, which alone holds the key, and deleted from the watch once the transfer finishes ^[19]. Max Weinbach wrote on X that the iPhone has had a Secure Exclave since the iPhone 16. He wrote too that an on-device model cuts the transcript to less than half its length, stripping filler words and anything that suggests tone ^[20]. The condensed text goes to Private Cloud Compute with calendar data, Now Playing information and broad location labels such as home, work or school. Apple's document says precise location and specific points of interest are left out ^[21]. Recaps disappear after seven days unless the wearer saves them ^[22]. Those protections all govern what survives the conversation, while consent under the statutes the attorneys cite is owed at the moment of capture ^[5]. Audio runs on secure hardware that the operating system, Apple and anyone else cannot reach, and since there is no recording, nobody can share, forward or hand over anything, the company said ^[24]. For whoever has to decide whether to switch this on, two questions decide it. Does the conversation fall under a state that needs everyone's agreement? Did the wearer say out loud that the watch is listening? Three of the four combinations are fine. The fourth, an all-party state and no announcement, is the one Blair calls a criminal violation ^[16].