Currently the legal definitions around the uses of power such as search and seizure rely upon determining if a situation had an expectation of privacy. Police need a warrant to search your home because most people expect the home to be "private".

One particularly interesting example of this Kyllo v United States[1], where the use of infrared cameras to find grow lights inside a home was ruled a search (and thus needed a warrant) because infrared cameras were not common so the general public did not in practice see infrared emissions as "available to the public". The opinion of the court, however, did point out that their decision would be defunct when infrared cameras became generally available and used by the public. At that point the public would no longer expect infrared emissions to be private. (Presumably this expectation might create a market for e.g. ir-blocking paint? If that's even possible?)

Installing microphones that regularly send data to unknown 3rd parties normalizes the expectation that spoken word in the home is no longer private. When perceptions change, so will warrant requirements.

> We're not living in a violent, single-party totalitarian regime.

If that's you're criteria for when to act, you are always going to be too late to do anything. Some of us prefer the idea of detecting problems early - hopefully before the become violent.

[1] https://en.wikipedia.org/wiki/Kyllo_v._United_States

> If that's you're criteria for when to act, you are always going to be too late to do anything. Some of us prefer the idea of detecting problems early - hopefully before the become violent.

Your argument is different to, and stronger than the parent's though. The parent's argument was not that we were on the thin end of a wedge to a totalitarian state, but that in some senses we had already passed that point because there was a voluntary element to it.

Still though, IR cameras by their nature are always-on and always transmitting data. They'd be pretty useless if they weren't! That's not the case for Google Home or any of the solutions currently on the market. Google Home also no-more violates this condition than say, a set top TV box that sends data on my TV viewing back to a central server.

Always-on isn't required. The deciding factor is what the general public expects.

If the public expects that a home might be recording their speech and sending it to a 3rd party - and thus theoretically has the opportunity to not speak or employ some sort of defense being recorded - then then Kyllo v United States suggests 4th Amendment protection no longer applies.

Just look at how many people here on HN already claim to no longer expect privacy.

> voluntary element

That's my point - your choice to voluntarily normalize surveillance impacts my 4th Amendment protections.

> Still though, IR cameras by their nature are always-on and always transmitting data.

In the case of near-IR night vision CCTV cameras in common use, yes. The case of Kyllo v. United States specifically involved very expensive long-wavelength thermal imagers that are not typically used in civilian CCTV applications [0]. Furthermore, federal agents specifically pointed a mobile camera at the house to see temperature changes behind the walls, which differs from a static always-on installation.

[0] There are fixed thermal surveillance imagers (see http://www.flir.com/surveillance/display/?id=64664 for FLIR's latest), but they're typically used for military installation security. Unlike NIR cameras they require no external illumination, and are tuned to detect the black-body radiation naturally emitted by warm-blooded things.

This is an interesting and relevant point. However, I believe that the expectation of privacy could still be arguably maintained because of the contractual relationship between the user and the service provider. The EULA and privacy policy would limit data sharing. There remains an expectation of privacy between two parties, and only within the context of the 'trigger command' that starts the process. A warrant obviously by-passes this, but you are talking about warrantless searches.

Our wonderful government disagrees.

https://en.wikipedia.org/wiki/Third-party_doctrine

Would that not still be different? The public doesn't have access to El Goog's recordings, so surely a warrant would still be needed as that precedent would not apply?

Google is part of the public. Yes, this would depend on the interpretation of the court, but the point is that the expectation of privacy no longer exists when you are explicitly sending that audio to a 3rd party.

I'd be very interested in how that would play out in a court. My gut instinct would be that there is a distinction between consenting to send voice data to a named company and conversations being deemed public. Of course, my profession is not in the legal domain.

No. You're just making things up. Law enforcement must have a valid and legal warrant to access Google voice logs.

I'm not talking about voice logs or anything else recorded by Google Home itself. Once people expect that their conversation can be recoded by a 3rd party, they no longer have the protections against other recordings being made.

> Law enforcement must have a valid and legal warrant

Kyllo v United States established a bright line test that the need for a warrant specifically depends on if the technology used is commonly available to and used by the general public.

Except that's not how any of these devices work. Nobody has any expectation that their words are being recorded until such time as they say "Okay google" or "Alexa".

This is about expectations. not technical features. If you walk into any random home, can you expect that your conversation is private? Today, that's already hard to answer because you first have to know if there are any Google, Amazon, Apple, or other devices that might start transmitting your conversation to a 3rd party. Do you know every possible wake word used by every devices with this kind of feature?

At some point it is impossible to know the list of possible words you cannot speak and the public will simply start assuming that they may be recorded anyway. When that perception changes, so will legal doctrines.

Not really, because every one of them gives an audiovisual prompt that speech is now being recorded. Nobody mutes those prompts because it makes the devices harder to use.

You can own every single smarthome device in existence today and quite rightly assume that you are not being passively recorded. These concerns amount to overblown FUD.

> Also, the 1984 comparisons just don't bear fruit. We're not living in a violent, single-party totalitarian regime.

... But we could be one day, and some people are. The transition to cloud listening devices in every house and the transition to totalitarianism each only need to happen once. The thing Orwell missed was that these devices could be made useful enough to consumers for that transition to be first.

It has been said many times before by commenters here on HN, and is worth repeating. Cloud-enabled listening devices already have saturated society and are nearly omnipresent. They have a form factor that fits in your pocket and also includes discrete surveillance-enabling conveniences such as: GPS tracking, always-on microphones, high resolution cameras, target-specific network graph data, and access to very large (and inconvenient to monitor) data storage.

If a person is concerned with the practical security implications of Amazon's Echo or Google's Home, I expect they should be terrified of the potential vectors for misuses that their mobile phone allows.

The future state of government regulation and the ability for those in power to use the digital tracks I've left do concern me, but I've found it very difficult to decide how to combat it without fully disconnecting and moving off-grid. There are ways to mask and obfuscate certain parts of that record, to be sure (Tor, proxies, not using Echo or Home, no connected devices, etc.) but I like the benefits that technology brings to me life and I don't really want to move into a remote, signal-free location to live a spartan life.

For sure. My argument is that, in general, consumers have found all of those listening devices more useful than terrifying.

I don't think Orwell imagined that universal surveillance would happen due to consumer demand rather than imposition by the totalitarian state.