If you leave your mobile accessible by voice commands without any kind of authentication, you'll always have issues.
Even just to changing the trigger phrase to something other than OK Google or Hey Siri! Should be enough to avoid issues.
University boffins have brewed one of the most complex mechanisms for loading malware onto phones by way of surreptitious Google Now and Siri voice commands hidden in YouTube videos. For the attack to work, phones need to be in a state where they can receive voice commands - a feature often left unlocked - and close enough to …
This post has been deleted by its author
"Sounds more like The Borg...
"We are the Borg. Lower your shields and surrender your ships. We will add your biological and technological distinctiveness to our own. Your culture will adapt to service us. Resistance is futile."
"Hang on lads, it's a Windows Phone !"
"Bugger. Nothing worth pinching from these Humans then".
Daleks will change their voice synthesiser to match the local language... So perhaps they are attacking a Klingon colony.
But "EXTERMINIEREN !" is not even a real German word - they should have been screaming "VERNICHTEN !"
Of course, they didn't quite get the voices right in the 80's
It sounds like you could happily include it in a techno track and it would not sound out of place. Now to get my brand of techno in to the play list for a public place, a railway station for example. No wait, announcers already sound like this, may be someone has beaten me to it. Another plan for world domination thwarted.
Not doubting the article, but it has just triggered a vague memory of some sort of scare ?? marketing ploy ?? to do with kids toys.
1980s ???
Wasn't there a kids toy robot thingy which "reacted" to audio codes in the (cunningly linked) TV cartoon ?
Or have I just given some startup an idea ?
"Wasn't there a kids toy robot thingy which "reacted" to audio codes in the (cunningly linked) TV cartoon ?"
It's called a smartphone.
They can embed signals within the passband of digital audio on TV that are too high frequency for most adults to hear.
Some Apps have the code to call the "mothership" so they know what you are watching. MUCH easier than voice recognition. Possibly Facebook App, it uses the microphone?
WiFi Barbie is a little scarier on the level of "Teddy always knows best" or whatever the Harry Harrison story was.
Phones (and some car radios) USED to have local voice recognition. Now it seems (on Android, Amazon Echo, Windows Phone and iOS) to be so called "cloud based". I have data mode on Mobile always disabled (to save money and for privacy) and wifi off by default (to save battery)
I could understand the words being spoken so I wonder why they bothered obfuscating them. The demo with the background noise didn't mask the words either as they were played MUCH LOUDER than the background burble. Let's see a demo that works in a busy pub on a Friday night.
I'm more worries about Radio 4 taking control of my phone. I'm sure John Humphrys never says 'OK Google' or 'Hey Cortana' at 7:45am but the Today programme keeps waking my phone(s) up.
This will totally jeopardize my plan to have a robot butler if it's going to be eating the TV every time Bart mentions his shorts or Bender mentions his shiny ass. Both those things obviously preferable to it killing all humans obviously.
Coat.. I'll be getting my own, the one with no robots in the pockets.
This post has been deleted by its author
That doesn't sound very obstructed to me, i could understand every example, notice how the test speakers they use are 2.1. As far as i can tell all they did is use a frequency shelf to squash/reduce the upper frequencies. From what i can work out they needed a reasonable low frequency speaker to make it work effectively, hence the 2.1. Finally most people i know dont set their device for always listening due to the battery sucking effects of having it always listening so its a bit pointless. Now if they had used some form of phasing effect passing different parts/components of the audio to left and right channels and use some stereo widening effect to deceive the human ear, even better doing this with a 7.1 system, but the sound would recombining in the mono microphone, i would have been more impressed, but instead they just fiddled with EQ and declared it a success.
The majority of users - your mum, kids for example, non-techies or those for whom security online is something nerds and the tinfoil hat brigade get upset about and they need to get a life (until someone hacks their phone and gets theirs) - they leave everything bog standard, unchanged and unprotected. How many people still have laptop mic and cameras or phone cameras uncovered despite well-published RAT attacks? This is purely a proof of concept system. The fact a distorted voice can be made to activate a phone even with background noise means that it could be refined so that the commands are HIDDEN in the background noise. So all kinds of havoc could be created. Or surveillance...
How many times have you stood on a railway platform or in a large atrium and not understood a word of the announcements? But suppose your phone can? Would you think it was a phone attack or 'just the usual c**p announcement system"?
Remember the TV advert for X-Box or Kinnect - can't remember which one it was - a year or 2 ago that activated peoples consoles of they were switched on and near the TV? Suppose they put something in the ad to activate your phone and pull up the website for the product? Or worse... order one for you? And here's a free virus while we're at it...
And who was it a few months ago on this hallowed site that said his phone was pulling up websites for sanitary pads after a conversation in the car with his wife?
Or maybe I'm just paranoid. Prepare to be Boarded lcon - cos you might be!
I'll get me tin hat.....