Slashdot Mirror
Samsung SmartTV Customers Warned Personal Conversations May Be Recorded
An anonymous reader writes Samsung's privacy policy includes details that its Smart TV voice recognition feature may pick up on personal conversations and transmit private communications to third parties. Buried in the privacy policy related to the smart television, Samsung advises users to be aware that any snippets of conversation might be captured by the software which allows them to control their television sets with a series of commands. Questions have been raised about who these third parties could be, what the information is used for, and how the data is being transmitted – with potentially unencrypted voice clips left exposed to hackers.
Generally, voice commands on phones require an activator. Such as "Siri," or "Okay, Google," or "Hey, Cortana." These are phrases that probably don't seep into the average person's life too often, so they're fairly safe to use as activators. I don't have experience with iOS or WP's apps, but Google Now requires you to "train" it to the sound and cadence of your spoken phrase, "Okay, Google." This allows the phone to detect the activator phone-side, without sending information to the mothership. It's also easy to test this, simply turn off network data (perhaps by putting the phone on Airplane Mode) and activate Google Now by saying, "Okay, Google." The screen will pop up, allow you to speak, but it won't be able to contact Google to parse your question. So it's clear that Google Now handles its activation phrase offline, and although I'm not certain, I can guess that other phones do similarly.
The other day I discovered a new iOS feature I had no idea existed. While sending a text to a customer, I hit the microphone icon by mistake. Another person in the (parked, idling) car was muttering about how the customer was a moron, which, although true enough, wasn't something I intended to include as part of the text. Not hard to see where this story is going, right? Well, it's even dumber than you're thinking.
I hit the 'Done' button to make the voice input UI go away, finished the text message, and hit 'Send.' As far as I knew, there was no danger that my friend's comment would be added to the text message. The car was too noisy and her voice was too low.for the speech recognition engine to understand, and in any event, the "moron" comment didn't appear in the outgoing text message. No problem.
Except what did appear in the text message, visible only after I sent it, was a small attachment balloon with a waveform. Apparently, iOS now sends the captured audio file as a binary attachment if it can't extract any recognizable speech.
So the obvious question is, what kind of drugs are these people taking? Is no one at a Fortune 500 company capable of thinking anything through these days? Do the programmers who think these features are "cool": and "awesome" not have managers with a three-digit IQ?
Fortunately for me, my phone lost its signal right about then, and I was able to kill the text app while it was still displaying "Sending." I knew from experience that iOS's text app didn't attempt to provide guaranteed delivery, and sure enough, when I restarted it, it had forgotten all about the message it was trying to send. So in a sense, I was saved by the same dumbshit programmers at Apple who tried to ruin my day.
It seems we have to adopt the same attitude around microphones that we normally apply to firearms. The gun is always loaded, and the mike is always live.