The latest Apple Watch models have some new features: they can listen to the sounds around you all day, transcribe snippets of conversation on demand and write up summaries of everything they heard.
The new features may run into problems with privacy laws, particularly those that require consent from people being recorded, but also with unspoken social conventions about speech.
When we talk with other people, we have some expectations about what happens to our words. Our comfort comes from what information scientist Helen Nissenbaum calls contextual integrity: the expectation that information moves around according to the norms of the setting in which it is shared.
What you say across a cafe table, in a bathroom, at a dinner party, at a town hall meeting or during a television broadcast is said for the people present and for the purposes apparent in that moment. Until now, our ambient conversations were not said for an auto-magical summarizer to fold into someone else’s searchable life records.
Apple can encrypt, minimize and delete all it wants, but data protection isn’t the problem. The problem is that a conversation has been quietly rerouted from the room where it happened into a device, service and product.
Did you read it? According to Apple it’s not recording. It’s also not transcribing, it’s just a summary.
You’re completely missing the point. You’re referring to recording as a noun but everyone else is referring to the verb, the act of recording. The very fact that it is capturing data for further processing regardless of how long it is stored or where is the problem.
It doesn’t matter. Wiretap laws, including federal and many state laws, don’t care if the recording is saved or even if an audio recording is created in the first place. Under the Federal Wiretap Act (18 U.S.C. § 2510(4)), an “interception” is defined as the “aural or other acquisition of the contents of any wire, electronic, or oral communication through the use of any electronic, mechanical, or other device.” Many state laws make it illegal the moment the audio is saved into an ephemeral buffer, which Apple does.
If the SoC is receiving a signal from the microphone and storing it in RAM at all, it’s fucking recording!
*facepalm
Not to you for being helpful and posting relevant info, but for Apple for being such ridiculous assholes.
It’s recording into the device memory and then transferring to the iPhone’s device memory. It’s still recording.
Also it says it does transcribe it. “On-device speech recognition transcribes the audio to text, and then an on-device language model generates a condensed version”
I think this is an option that a user can enable. So it’s not something apple is doing without the users knowledge. Still, other people may not be aware they are being recorded and that the person is getting summaries of the conversation.
I’m sure there are audio recording apps available for iWatch which could be used more nefariously. Still, I’m not sure it would be legal for a person to create and use a device that “secretly” transcribes audio, but it’s Apple so I’m sure their lawyers can figure it out.
Depends on the state. Some states require everyone involved to know they’re being recorded, some only require single participant consent.
Can you just read the document in full? It’s not that long. I don’t want to quote the whole thing.
It doesn’t record to device memory, it’s a continuously overwritten stream of data into the Secure Enclave.
This this absolutely nothing like what an off the shelf recording app would do, and it would be impossible for an app to access any part of this pipeline.
Yes technically there is a transcription step in this pipeline, as you referenced. But again, this is a continuously overwritten buffer inside of a black box where the output is a high level summary.
Yeah it’s in the secure enclave RAM while it’s processed. RAM is memory.
In my state it is illegal to transcribe and summarize a conversation if both parties do not consent. Saying that it records and transcribes and summarizes really quickly doesn’t seem like a valid defence.
This is just trying to find another edge case legal loophole. It’s creating a summary of every conversation; therefore it’s recording, by the very definition of the word. Just because it’s not a word for word transcription doesn’t mean it’s not a record.
Just because every word wasn’t written down doesn’t mean I’m ok with it.