Win Back Some Privacy With A Cone Of Silence For Your Smart Speaker

To quote the greatest philosopher of the 20th century: “The future ain’t what it used to be.” Take personal assistants such as Amazon Echo and Google Home. When first predicted by sci-fi writers, the idea of instant access to the sum total of human knowledge with a few utterances seemed like a no-brainer; who wouldn’t want that? But now that such things are a reality, having something listening to you all the time and potentially reporting everything it hears back to some faceless corporate monolith is unnerving, to say the least.

There’s a fix for that, though, with this cone of silence for your smart speaker. Dubbed “Project Alias” by [BjørnKarmann], the device consists of a Raspberry Pi with a couple of microphones and speakers inside a 3D-printed case. The Pi is programmed to emit white noise from its speakers directly into the microphones of the Echo or Home over which it sits, masking out the sounds in the room while simultaneously listening for a hot-word. It then mutes the white noise, plays a clip of either “Hey Google” or “Alexa” to wake the device up, and then business proceeds as usual. The bonus here is that the hot-word is customizable, so that in addition to winning back a measure of privacy, all the [Alexas] in your life can get their names back too. The video below shows people interacting with devices named [Doris], [Marvin], [Petey], and for some reason, [Milkshake].

We really like this idea, and the fact that no modifications are needed to the smart speaker is pretty slick, as is the fact that with a few simple changes to the code and the print files it can be used with any smart speaker. And some degree of privacy from the AI that we know is always listening through these things is no small comfort either.

81 thoughts on “Win Back Some Privacy With A Cone Of Silence For Your Smart Speaker

        1. Smart-e appears to be something less than Alias, as it only UN-mutes with a physical button push.

          Not sure I have the skills to build myself, is there a commercial version that responds to “Alias” device wake word?

        1. LJ – actually in any military weapons fire control system scenario you’d never have A.I. take over human control of a kill system. Voice control should be relegated to low level commands not kill systems. Dr Stephen Hawking mentioned that before his death I think.

          In Star trek, ship self-destruct has a 2-party executive authentication. The same applies to nuclear release in US NAVY, except now it’s more than POTUS, now the JCS must concur (since POTUS 37),

          Raising shields should be a low level AI command and be automatic in some scenarios.Well not so much with the USN Phalanx System. You definitely don’t want that ship shield automatically coming online. In the ST-TOS they learned not to let the M5 (AI System) take total control.

      1. Alexander Weiss – That’s because Star Fleet’s IT division designed the Computer’s directed speech recognition to listen to background conversations beforehand to know when the context is directing to the computer rather than just falsing on random hits on the verbal command being heard. That’s actually a science in new speech recognition today. That’s what Alexa is kinda doing too but is really getting the paranoid types to complain too much. Trust me if THEY want to hear what you are talking about they will find a way (Remember Gene Hackman in The Conversation? And he was playing a real life surveillance expert.)

      1. Actually, getting it to trigger on “Sarah” and then say to the echo “Alexa, ask Sarah to turn on the TV” (of course it’d require a custom skill) and then repeat whatever else you say after that would result in something like “Sarah, turn on the TV.”

      1. Is there actual evidence of the companies making these devices stealing data from people? Genuinely asking. I own three echo dots and use them for controlling my lights and for timers and the like.

        1. Stealing data no, but they do collect it and there have been instances of that data being sent to the wrong person. for reference: https://gizmodo.com/the-amazon-alexa-eavesdropping-nightmare-came-true-1831231490

          If they have your data in storage, the questions becomes what could they do with that data? It doesn’t matter what they are doing with it now, because they have it and once they have it they can do anything they want with it, Including re-scanning it down the road for other triggers or even just selling it to a third party who could then do whatever else they want with it.

        2. No its not about cyber-theft, its data slurping. Google, Microsoft, Amazon etc. is only collecting information on you by your background conversations detected via mics and maybe even cams. They admit to it and say your opting in to it somehow. Google listens to everything you do. They claim they don’t archive it by your name but I wonder how they know what ads to push to you specifically. Jeff Bezos claims he deletes all that Echo data once they get what they need, your ecommerce preferences. However, NSA claims they do the same thing, dump the data once they realize you are an American citizen and should never have been a POI (person of interest). So its all about faith that someone is doing their due diligence to comply with FISA court. Jeff is not under FISA so he can do what he wants with his Echo DOT archive, but he claims its dumped almost immediately. Is he telling the truth? I think so.

          Put black electrical tape over your cam. And put dummy plug for your PC/Laptop mic. Echo dot also can take the same mic muting plug. Someone told FBI DIR James Comey to do this. So maybe good advice?

    1. Yeah, this is one of the silliest computer to computer interfaces. Just throw out the smart assistant, nobody gets any important use out of them anyway. Everyone I know with one has gotten bored with theirs and stopped talking to it. It’s so gimmicky. I’d love to have meaningful control over a computer with my voice I guess, but this isn’t it.

  1. Great innovation.

    https://www.youtube.com/watch?v=HWtPPWi6OMQ
    CONE OF SILENCE (GET SMART TV SHOW 1965–69)

    One way to protect you from Google/Alexa snoops is when you have a Magenta alert and a Class A, scenario is to immediately stop talking, press the button on this gadget in my next posting which is $19.95 USD at Home Depot, and continue your sensitive verbal traffic. The assistant is audio blind. Or find a way to repeat your room audio in the 3-6khz, pass-band which is the blocking band-pass for Amazon Echo to ignore audio from TV Amazon adverts (i.e. during Superbowl) which says the word “Alexa”)

  2. I guess the Home Depot gadget image is being blocked by OP for some reason. Its called the Big Red Rooster personal care sound projector. its $20 bucks. You can look it up at Home Depot dot com. You’d have to put the Big Red Rooster on top of your Alexa Echo. It has a WHITE NOISE push button toggle momentary action switch on front panel. Its a white box and it looks very techie too.

      1. Alexa can play “white noise” too but how can you trust she’s not using audio anti-phasing to try and filter or mitigate white noise? Its easier when the source is a common synchronized source. Try this 10-hour endless loop on YouTube. https://www.youtube.com/watch?v=jRyp5SvXjpI
        Make your own HTML page that activates it or stops it with a large HTML button marked [CONE OF SILENCE].

        Listening to it now. It really blocks room audio quite well, and it has therapeutic benefits too. You could also use text to speech or pre-recorded .wav files pre-canned menu-driven Alexa commands in your HTML app too that interrupted white noise temporarily to tell Alexa to do something.

        1. Wade… not me Wade. I love the thing. I just put on Rush Limbaugh or Fox News and let Alexa listen to total disinformation misinformation (sarcasm)

          I’m just an idea man here on HaD. I trust the gadget and could care less. Unplugging it is not an option as this clever gadget is uber-cool. I can live with the cyber-snoop as its not logical to think this thing is like Echelon et al. It would be tantamount to doing this:

          https://encrypted-tbn0.gstatic.com/images?q=tbn:ANd9GcSVNeENO1PG9GBCRRccQsn9y07mUvt3pQRQtuCbT9tbD34zLcHu7A

          However, cyber-tracking by Google, Amazon, and other eCommerce cyber targeting cookie monsters is pissing me off. Its too spooky. Too many coincidences.

          Mr. NSA (Snowden) hasn’t vilified or hacked Alexa yet… but he does warn that who knows what Amazon (et al) is doing with our Cloud ware. Echo DOT (et al) is a weak threat by eCommerce not the 5-eyes group. They have much stronger threats you probably never heard of YET. Watch the new NETFLIX series Pine Gap to get a glimmer of what “THEY” can really do..

      2. Why not just pull the plug of the Echo device? Or is it battery powered and wireless? Then the best is to place it into a bucket of water. That should mute the sound sufficiently and prevent further listening attacks.

        1. Martin – the point being the Echo Dots are actually useful like some Greek mythological 9 muses, so no I say save it but cripple it a little. Pumping wide band toggle-able white noise seems workable. And the snoops are just benign ecommerce companies like Amazon and Google trying to learn your commerce preferences for future targeted ads. The alphabet soup like no such agency ALREADY has much more nefarious means to violate your privacy- per Snowden, this ain’t one of them. Your standard tango (aka bad guy) doesn’t go out and buy an echo dot. Kinda’ defeats the whole bad guy schema? But I would worry about the boys in red square co-opting Jeff Bezos (Echo DOTS) the way they did Zuckerburg. They could use the Echo Dot data to build psych profiles on VIP users. However, Jeff insists his Echo dot data center is uber-secure, but that really doesn’t mean much with people like Snowden and his new Vodka drinking buddies hacking everything to pay their rent in Moscow.

  3. A partly unrolled French safe would be more appropriate, as this is prophylactic for those thingies. George Orwell is spinning in his grave at 7200 rpm. All in all a good device.

  4. Switch it off, take it out in to a field, dig a hole, bury it deep. Problem solved.
    or Alexa — what is the meaning of life the universe and everything? I’ll wait.. seven and a half million years you say.. yup. no problem, I’m very patient, I can wait.. ‘click’ … silence…

    1. “The answer is 42.”

      And after asking that, everybody I know who has one stops asking it important questions and just uses it as a disk jockey. In olden days we had to walk across the room to change the radio, but no more.

      Half the info in the internet is bullshit, anyway.

    1. You can easily produce the effect without a 3D printer: Use a can of polyurethane foam and embed the device into it’s sprayed out contents. Should have sufficient sound damping.

  5. Do you americans really believe that Alexa or Google gives you access to “the sum total of human knowledge” ?
    Because if it’s the case, I really pity you.

    BTW why would a interface with, at max, 150 bits/s (human speech estimated entropy, see https://onlinelibrary.wiley.com/doi/pdf/10.1002/scj.1076 ), would be faster/more ergonomic than a keyboard or a mouse or my fingers movements and a visual feedback (which gives many order of magnitude more data per second) ?

    The only reason I’d agree with is that it requires zero effort to learn to use. But if this is the reason for the “assistant” success, we are doomed.

      1. Sure, but how much info is there in speech ? Try to read aloud a paragraph from Alice in Wonderland and notice the time it takes for you to do so. Then just read it in your mind, you’ll see you’ll be 10x faster (at least).

        Because speech is a low bandwidth channel to transmit information with a high error rate (around 5%, usually more), unlike something displayed. I’m almost sure that if you had to describe aloud all the actions you’re doing on a keyboard (locate key ‘Z’, press key ‘Z’, release key ‘Z’ and so on), you’ll see speech is much slower. Just try to control a fortnite player with speech, you’ll see it’s impossible to do so (you are much faster with your keyboard).

        In the end, using “text” to control something is almost every time the wrong method.
        Speech to text to order to action to text to speech (what alexa/ga is doing) is wrong, compared to me flipping a switch to shut down the light. By the time she’s done, I’ve moved my .ss to the switch.

      2. Have you tried unplugging it and then *not* plugging it in again?

        Seriously – boot time on Alexa is a few seconds. Just unplug it – no internal battery. If you’re worried, wrap it up in a towel…which you should always bring since they’re massively useful.

  6. Why use white noise. How about containing the top of the unit in a soundproof box and letting the pi pass your audio command through to a speaker above the mic. Are these things sensitive enough that they can pick up audio vibrations through their cases?

    1. ballooner – actually that’s a good idea. If somehow you could make the pi pass only speech in the 3,000 to 6,000 Hz command ignore band pass built into the Alexa units. They do that to stop command falsing during Amazon adverts. Other companies are using that too. Local News shows haven’t caught up yet. Sometimes they trigger the Alexa “my briefing” mode and she goes temporarily deaf. (Put that doesn’t help the paranoid types – still need to figure out how to preserve the desired voice command)

      And YES the units have a very unique sound detection scheme. It discriminates other sounds out too. So it can hear you within a cacophony. And it learns to recognize your voice over others.

      White news audio jamming is a typical anti-surveillance technique. It seems to be effective but it has its drawbacks too.

  7. So he built an Echo to go on an Echo? The wake word is already customizable. And prior to the wake word, the Echo only records a few seconds of speech into a local buffer that is constantly re-written. Sounds like this device functions the same way??

  8. This spying devices remind me of Dragon Naturally Speaking, the voice-recognition add-on for MS Word of the 90s. It sounded like a good idea until I realized that my own foreign accent made the software useless.

    Why in the world would I want these trojan horses in my home, which can later be used to violate my civil rights? You think I’m exaggerating? Ask the dude whose Alexa was used to prove the guy was planning to kill his victim. Buh-bye, fifth amendment.

Leave a Reply to ljCancel reply

Please be kind and respectful to help make the comments section excellent. (Comment Policy)

This site uses Akismet to reduce spam. Learn how your comment data is processed.