Hacker News Reader

Comment by Multicomp | Hacker News Reader

I must be dumb but in today's world of the big 4 spying on you all day long, I'm trying to kick out voice assistants, not bring more in. After Google Now, my voice control needs are done and someone said enough is as good as a feast.

Now when you get Mrs. Roddenberry's TTS voice powering a smart home, we may have to revisit.

View on HN

These devices all ride on your own personal network. Their traffic is easy to monitor for timing and size of the data transmitted outward. If you are worried about spying you might put a piece of tape over your webcam, but I can't say I have heard of anyone physically disabling their laptop's microphone. These voice assistants are basically the same security threat as those open microphones on your laptop. Both of them are still less dangerous than a smartphone that has its own network connection and can more easily hide its spying.

Many people physically disable their laptops microphone. Some people in my office do and I think the most famous case is Zuckerberg.

https://boygeniusreport.files.wordpress.com/2016/06/mark-zuc...

The difference being someone might actually care about what Zuckerberg is saying.

And? I was simply refuting the comment "..I can't say I have heard of anyone physically disabling their laptop's microphone". I know people who do this, not just Zuckerberg.

Putting tape over a microphone doesn't physically disable it. People have been able to get audio out of a room by watching the vibrating of windows or simulate a key logger by listening to the sounds a keyboard makes. If you are worried about being actively spied upon, nothing short of physically disconnecting the microphone is 100% secure.

Just because you can monitor the traffic doesn't mean you can verify what it's sending. If the assistant is designed properly, it'll be using pinned certs and everything it can do to encrypt and compress the data, but to be usable it's probably listening and throwing away a lot of data to be responsive (like how the new iPhone cameras constantly capture when on, and give you appropriate moment right before you press). There is a possibility that additional trigger phrases may exist that Amazon/Google/etc capture. It might be rare enough to be difficult to trace.

I wonder what data was handed over in that Echo subpoena for the murder case? http://www.cnn.com/2017/03/07/tech/amazon-echo-alexa-bentonv...

Yes, you can't tell what it being transferred. But knowing when and how much it is transferring shows that it isn't actively listening to everything.

If you are worried about companies like Amazon and Google lying about what these devices use as wake words, why would you trust Google, Apple, or Microsoft that they don't lie about the same thing regarding your phone or laptop? That is my main problem with these arguments regarding voice assistants. These devices are really just a microphone and an internet connection. They are no more a tool of Big Brother than your phone, computer, TV, Xbox, or any other device with a microphone and internet connection. In fact, many of those devices can do a better job of hiding their compromised status.

There is an option: Don't use voice assistants.

That's my option. I trust Apple (and Google for the few devices in my home) to not be listening persistently when they say they don't. I don't enable Hey Google, Hey Siri. That could be naive but as you said you have to trust at some point (even if it's the compiler or hardware).

I can disable hardware at the USB bus level on my MacBooks and other generic OS - I don't use microphones on them anyway. If you really distrust you can wire a headphone in and disable that.

Persistent voice recognition is a rubicon I don't want crossed in my home. It's a "feature" I don't think is needed. DONOTWANT is a signal I want to send to the market, fwiw, by not buying these devices.

Unless you also keep cell phones out of your house, this really seems like closing the barn door after the horse is looooong gone.

I think the difference is that those devices were created and used with the expectation that they won't monitor users. The new age devices primarily want to monitor devices.

What’s the difference? Devices like the Echo sit idle, using local hardware to listen for a specific wake phrase. Upon hearing it, they activate, send what you say off to the cloud, and take action accordingly. Modern smartphones do exactly the same thing. In both cases there is the expectation that they do not do anything about what they hear besides commands prefixed with the wake phrase, and in both cases you pretty much have to take the manufacturer’s word that this is actually the case.

People always jump up and defend IoT devices, and while they make reasonable points (you can monitor their bandwidth, you already have a smartphone, etc), I'm just not willing to provide yet another potential entry point into my personal life, especially one that provides so little added convenience over the smartphone that is always within reach anyway.

i’m not here defend it, I’m just here to tell you that I’m sick and tired of people whining about it every time there’s an article like this.

Simply don’t use it if you don’t want it. Move on. Quit reading stories like this.

[Update] I just blogged my comments.

https://h4labs.wordpress.com/2017/09/27/groundhog-day-amazon...

The point is that I heard you the first ten times.

At this point, that's kind of like telling non-smokers to just not breathe smoky air when they're around other people. These things are everywhere, people carry them in their pockets, millions of households contain standalone units. You're more likely to be in listening distance of a virtual assistant than not.

Reply on news.ycombinator.com