Is your smartphone listening to you? | Hacker News Reader

Is your smartphone listening to you?

bbc.co.uk

130 pointsCarolineW119 comments

The article takes it into a boring direction, and they didn't mention Samsung who is most active in this field. Vendors bypass appstores with their proprietary apps. Once upon a time I had a Samsung phone, and when I configured its email app, I saw Samsung's own servers, and not my phone, logging in to my email server. Which means they stole my password. That's when I tossed out Samsung devices once and for all.

But anyway, here's the courtesy tl;dr of the article:

Can it be done? Absolutely. Will it be done? Most likely. Have many people suggested that it's already happening to them? For sure, all over the place.

Is it allowed by Google, Facebook and the like? Most definitely not. So is it happening? Probably not, it's just a coincidence that you talk about something specific, and then suddenly find all kinds of ads about it.

Here's a dictionary definition of 'coincidence' followed by an explanation from a professor about how our minds have evolved to correlate stuff all over the place and see patterns in them, even though they're just coincidences.

> Once upon a time I had a Samsung phone, and when I configured its email app, I saw Samsung's own servers, and not my phone, logging in to my email server. Which means they stole my password.

Would love to see some traffic logs regarding this.

Plus a copy of all the EULAs you consented to with Samsung when you used their phone and software.

However relevant this possibly is a court case it is less relevant when you are about to pick between a nexus, an iPhone and a Samsung S-series.

Also while IANAL, note that in a number of jurisdictions EULAs like the ones you seem to think about might not be valid. Norwegians consumer authorities recently mocked them and I applaud them for it.

With a Nexus you're trusting Google with a lot more info than Samsung could hope to obtain. So if this isn't about legality then I'm not sure what it is. Presumably Samsung aren't storing your email credentials so they can launch a spam campaign, and you've not provided anything like enough information to deduce that they're doing anything other than, perhaps, keeping the email moving.

I never read EULAs and I can imagine they're not valid everywhere, but they're useful as a measure of the intent of the company producing them. I imagine Samsung, for example, reserves the right to use any information you provide to ensure a certain quality of service. I'm sure Google have run their EULAs - valid or not - past similar lawyers.

Google has shown that they fire sysadmins who put their nose into customer data even though tje sysadmin in question had good intentions. AFAIK they also have a fairly strong security record. I'm not saying Samsung doesn't but I don't know.

It's also an absolutely wonderful blame for people so inclined to inform on others.

You set it up with your own IMAP(TLS) and smtpd for email, and in your server logs the logins were coming from Samsung IPs, not the wifi or cellular network your phone was connected to? Where were these Samsung IPs in terms of physical location and netblock? If they're doing MITM on their customers' traffic I'm really curious where it's going through.

Correct. The logins came from Samsung IPs. I wrote up a fat warning on xda but am unable to find it. Sadly it was too long ago for me to find it back.

If a vendor provided a physical switch with which I could turn off the mic, that would be a feature that I at least would consider very attractive.

After Snowden I just took a pliers to my laptop and ripped out the mic and webcam. It felt great.

Speakers can be used as microphones....

Very very shitty microphones, so the statistical risk is significantly reduced! And the hardware might not support doing that switch in software.

Actually, you can get pretty good quality (You could easily do speech-to-text on the produced audio)... If you want to do a simple test, plug-in your headphones into your microphone jack and start recording away. I just did it why my current cheap Amazon basics headphones and I can clearly hear my speech back. Also take a look at https://www.youtube.com/watch?v=GqNrLfbwAH8, I think that this shows you how to record from your laptop's speakers in Windows using Audacity.

True.

The video: No, it shows how to 'record what you hear' (windows mixer) nothing to do with the speakers.

I would claim that it is impossible to record from speakers that are connected to low impedance running output amps and record from them at the same time.

So if the speakers are playing music they are not recording your sing-a-long also.

Your solution is more foolproof, but Micro Snitch[1] is worth looking at - same company behind Little Snitch.

[1] https://www.obdev.at/products/microsnitch/index.html

That's in my proposal for a secure or private phone that I drop on people trying to build one. Hopefully it will get in a successful product at some point. Need one for mic, front camera, rear camera, and radios. Might compromise to have a physical switch for mic and radios with something inside phone (eg a jumper) to turn off others. I want to be sure I can cut actual power to mic and wireless, though.

If your threat model concerns with hardware/firmware exploits, then you can't trust the physical switches unless they cut all the wires - and unless you can validate that this is what they do. Theoretically speaking, a malicious (or threatened/persuaded/forced) manufacturer may implement a non-standard way to power the supposedly disconnected component using still-connected wiring, to be used for a special occasions.

/tinfoil

You should see my full write-up:

https://news.ycombinator.com/item?id=10906999

I went into detail on what it takes on hardware side:

https://news.ycombinator.com/item?id=10468624

Agreed, which is why I believe sound proof cases are the only sensible approach, and work with all current consumer devices.

This is why I got a shutter on m laptop camera. It's mechanicl, and I can see it's on or off. It should be a legal obligation to provide those, for camera and mic.

Regarding "legal obligation": in some countries the government may be the second main user of your camera and mic - next to you...

This is sadly true.

I want a DPDT switch that sits between the mic and the rest of the phone. That seems do-able to me.

I use masking tape over the camera, so I'm looking for a similar microphone solution.

(I don't want it inside the case, since I want to be able to flip the switch conveniently. Likewise, removing the battery is not what I want.)

Well, I just.. disconnected the mic in my laptop, the day it arrived (there is no webcam). As phones are concerned, I use my trusty SE ELM featurephone, so I'm not very worried about Googles, Facebooks, Samsunsungs...

I'd also buy such a phone. The switches to toggle each radio and sensor could be behind the removable back panel. I wish it were possible to easily disable in software mobile service without also disabling wifi and bluetooth as well.

At least on my Huawei P6 i can set it to flight mode and then turn back on wifi and bluetooth. Actually i have used the thing more like a PDA than a phone, as i have a well functioning HSPA capable featurephone that i pair up if ever i need data on the go.

> i have a well functioning HSPA capable featurephone

Make and model, please? I've been searching for something like that for a while now. Everything I've found is either carrier-specific, doesn't allow tethering, or is 2G only.

Long since discontinued, sadly. Its a SonyEricsson C702.

You mean it's a soft-switch that tells you it does that. You have no idea if the radio is actually running at any time or not. That plus their reliability issues are why I don't like soft-switches.

Well in my case there is no sim in the slot, so i can't say i care.

It still has an ID (IMEI).

That doesn't stop it communicating with the mobile network. At least in the UK, you can still call emergency numbers (999, Childline, Samaritans(?)) without a SIM.

>That doesn't stop it communicating with the mobile network. At least in the UK, you can still call emergency numbers (999, Childline, Samaritans(?)) without a SIM.

You haven't been able to do this since 2009. It was changed due to the high volume of hoax calls.

Huh. Someone should tell my phone; it still displays 'emergency calls only' when it is unable to connect to my own provider's towers.

>Huh. Someone should tell my phone; it still displays 'emergency calls only' when it is unable to connect to my own provider's towers.

That's emergency roaming. You can call emergency numbers provided there's signal from any network and you have a SIM installed.

If your are equip like me with a EMF meter, then its easy to confirm what your phone is doing when in airplane mode, with normally is nothing. A mobile phone is after all just a computer with a radio modem. It can't do shit without sending out RF.

I would recommend something like the 'Cornet ED78S RF Meter' to every technician with the tiniest amount of AU in circulation.

The reason I got it was because I have become electro-sensitive, and that's almost a death sentence in the modern city.

Those pictures of Stallman sitting on a mountain top and not a mobile phone mast in sight looked very appealing...

My phone LG-L80 99£ is GSM only, all else is off, apps downloaded on wired PC and installed via ADB interface.

Currently used as book reader for C++11 fourth edition and as Wikipadia offline reader(Kiwix) 12GB on SD card - music/mp4 player and the battery last for a week or more.

I feel I'm in control of my phone, don't know if that is justified or not?

What frequencies can an EMF meter pick up? One of my own designs of the past was to hide something in a cable to amplify a signal along 10GHz optionally with a beam hitting it. None of the WiFi security monitors looked for 10GHz. Neither did most cheap, spectrum kits.

The dedicated chips in phones with the radios could have extra functionality for other spectrum activated by a trigger, possibly received wireless signal. So, it's a concern to me.

The one I suggest goes to 6 Ghz

I can't imagine that pulsed digital transmission on 10Ghz would not create noise in the lower harmonics bands at all. The meter also have a sound mode where you eg. can hear the ~8-10Khz modulation of the mobile ~700/900/1900/2100Mhz signal.

You can clearly hear the difference between different transmission types as DECT phones, Wifi, GSM/LTE, and EM noise in general(PC's/HDMI cables/USB3 HD docks)

My central room heater has a small CR battry powered computer (that I had forgot all about) that calculate the bill and transmit the results every 2 minutes on 2.4Ghz

it makes a tiny ~0.5 sec pulse on the display/audio, and I thought for an hour that I had a hidden transmitter in my apartment, since turning the mains fuse off had no effect.

I had to clear the room of all tech to track down the source, felt a bit stupid when I finally found it.

So going down this road has some up's and down's vs staying totally oblivious of ones electromagnetic environment.

Appreciate the extra info. Might look into getting one. Lmao on the mystery of spying heater.

True, though with airplane mode they might be skirting the law as it now touches the whole interfering with FAA requirements are. Also it would seem it would be easy to verify for anyone who opens the case what the radio does.

That's not how subversion works. The subverted system will in fact turn off wireless in airplane mode. It's turned back on by a trigger. Stays on long enough to do the job. The commenters imply they would be using it outside an airplane. So, FAA wouldn't even notice.

Sell a micless phone. Actually, it's a tablet, and if you want it to go into surveillance / phone mode, you plug in your own goddamend mic. Most headsets have one.

Alternatively: a phone whose "off" mode is a battery eject.

Re battery eject: How can you be sure that the removable battery is the only one?

Fair point. Proofs are hard, but on a mass-market device, considerable secondary storage would be fairly evident in a teardown.

Microcurrent devices fed by a small cap or disguised secondary battery might function. But you'd have very limited capabilities.

You might well ask how sure you are that there's not a secondary microphone on the device, to which I can only suggest you cannot be too paranoid.

Brilliant. Never even occurred to me. People often use headsets, wired or wireless, for the mic. Can just not enable audio unless one is plugged in. That's so easy to verify. Thanks for stating the obvious to my overly-complex mind. :)

I've always found built-in/integrated cameras/mics to represent tremendous negative value.

1. I cannot unambiguously disable them.

2. I cannot upgrade / swap them readily for something better.

Aftermarket/detachable just seems such an obvious alternative.

And yeah, a simpler solution to a hard problem ;-)

See slide 30 on this Neo900 presentation: http://neo900.org/stuff/ohsw2014/ohsw2014.pdf

Also see the various kill switches on the block diagram (GPS kill, modem monitor, mic not directly attached to modem (so you could probably send fake audio/silence),etc: http://neo900.org/stuff/block-diagrams/neo900/neo900.html

See slides 16 through 23 here for some details on the modem/audio monitoring/separation: http://neo900.org/stuff/cccamp15/ccc2015talk/talk.pdf

Would this meet some of your criteria? If there is something missing you can discuss with the people involved (on IRC usually): http://neo900.org/contact

It meets some of my criteria. They're doing good work. Gotta ditch that TI processor when possible. Need some hardware and software modifications. However, actual switches for power and audio coming through Linux part are positives. Removes some low-hanging fruit.

> my proposal for a secure or private phone that I drop on people trying to build one

Can we see it?

Here's my write-up on securing Tor that shows issues and mitigations from high-assurance security in action with my own style:

https://www.schneier.com/blog/archives/2014/09/identifying_d...

Here's one write-up of similar methods applied to mobile:

https://www.schneier.com/blog/archives/2014/06/blackphone.ht...

Here's some links to HW and SW techniques that might be used in SOC to knock out the 0-days:

https://news.ycombinator.com/item?id=11848132

Note: You can see difference between few in real security on HW side. Mainstream, faux security immediately says IOMMU, stateless, whatever. High-assurance, real security is concerned with segmentation, pointer protection, making memory untrusted, TCB effects, optionally timing of ISA operations, and so on. Apps + their CPUs get hit all over. Securing one takes more than a mere IOMMU or something. ;)

Common workarounds include disconnectable devices like USB microphones, or creating your own hardware shields like electrical tape over a laptop webcam.

This is more a plug for the EFF than it is for a product, but these camera stickers have served me well and helped strike up a healthy conversation about privacy - https://supporters.eff.org/shop/laptop-camera-cover-set

might be a future option for Google's project Ara.

I'm not seeing that. Certainly I'd expect their basic phone to come with a mic. I want a way to physically disconnect an existing mic, not to connect a new one. Am I off here (probably ..)?

The Ara is a modular smartphone system. Since there's a clear benefit to being able to remove the audio/video/wireless modules, I would hope that they are fully replaceable and separate from the screen and CPU modules.

I figured they could even make a phone without any microphone and then let users plug in a headset when they want to make calls.

Yes, but maybe this is too ambitious. Maybe, for now, we should be happy if we can remove the battery.

The only real answer to this with current consumer devices is to have a soundproof case for your phone. Radio connection can be kept alive, and if the case were smart it could have its own internal mic, wired to an led or speaker on the outside of the case to alert the user that the phone is ringing.

It's a product I've been kicking around producing for a while now. Is there demand? Would you purchase for peace of mind?

I have come to peace with ubiquitous and comprehensive data collection of everything I send unencrypted over the wire. In a way it fits with current 4th ammendment ideas concerning zero expectation of privacy.

But just like the Supreme Court I strongly draw the line at the walls and shade drawn windows of my home. I do not want internal surveillance of my personal conversations with friends or lovers in my own home, and every internet connected microphone you own, in your macbook, Fire TV remote or cell phone can be almost trivially turned into a bug by any actor with the right exploit.

We've already seen it done by IT at a school district [1], school staff attempted to use laptop webcam photos of students in their bedroom to discipline them for drug use.

[1] https://en.wikipedia.org/wiki/Robbins_v._Lower_Merion_School...

You can also remove the microphone and use an external microphone (Bluetooth or headphones) when you need it.

I've thought about making something like this myself and I would definitely purchase one if it were available.

I wonder how to make an acoustically insulating phone case that is not so bulky one has to keep the phone in a bag. Acoustic insulating materials I've seen are fairly massive or have large spatial extent (or both).

This is why I use an app called DVasive which tells me when my mic or camera is in use. It's pretty good at figuring this out it seems, although I haven't found any nefarious apps yet that use my mic or camera unexpectedly (I've been using the app for a couple of years). It probably wouldn't work for vendors or other rogue ways to get access to my phones devices however.

How can you trust DVasive isn't itself malicious?

Isn't DVasive software founded by John McAfee? I'd trust it based purely off his personality. He's kinda crazy (in a good or bad way, depends on who you are), and he'd probably make a lot of noise if he found out his app was malicious.

Thats all is based on if he is somehow still involved with the technical details enough to figure it out. Most likely whoever put it in won't be talking about it to him.

It doesn't require any device privileges, and I can tell it's working just by opening up my camera app or snapchat. Other than that, I don't know.

And that it actually works?

In defence of the BBC, they are particularly sensitive to this atm. It is not unusual for journalists in the UK to be asked to remove batteries from phones, particularly when discussing certain international matters that may interest certain people. The concept of a smartphone being turned into a listening device isn't unheard of in the UK. For whatever reason, US/Canadian journalists are more trusting. I trace this to the aftermath of the london riots and RIM revelations. So I am not surprised to see the BBC publish something in this area more readily than others.

What do the journalists do when the battery is non-removable? Do they just leave the phones "far enough away"?

Not joking: I've seen them wrapped up in layers of tinfoil.

That's fine, if it's not recording to send later when defoiled.

Better off not taking it or putting it in another room, next to a radio tuned to static.

Mind you, when I'm pocket dialled I can rarely hear any conversations clearly. Doesn't take much to muffle audio input.

Unless the phone isn't simply recording, but is harvesting the mac address (or other identifier) of nearby phones in order to expose who the reporter is meeting with. Then the microphone is irrelevant.

I wouldn't be surprised if your smartphone was listening to you, but sadly these particular feats would be feasible simply based on all the information these people have been willingly sharing with google/facebook/etc.

If a lot of her friends/family are searching for "Motorbike accident, Thailand", it seems likely google would suggest it to her.

Similarly, if someone's boyfriend mentions to a third party on facebook something along the lines of "sorry we missed the <event> last night, @girlfriend was having a migraine", it's not so surprising if said gf is shown migraine ads soon after.

Bottom line is, I think people vastly underestimate how much private information they are willingly sharing. Sadly, these companies don't even need to listen in sneakily these days.

I have to say, I'm very sceptical about all of this, but one day a friend showed me a product (some speakers that use the table surface for resonance) and the ad showed a couple of minutes after on Facebook.

I started thinking on how devices can emit audio signatures for other devices around it to be aware of its presence. A couple of years after, Chromecast uses audio for authentication of nearby devices.

I honestly dismissed my paranoia all of these years with the argument that it would show in battery and data usage. But with offline language models, I'm not so sure anymore.

One theory I have is that advertisers keep track of IP address. So if your friend searched for something on your internet, it would show ads to you.

I know this happens because a few years ago there was a discussion about how advertisements were leaking christmas present purchases through something like that.

But what's really creepy is the article claimed 24/7 speech recognition didn't drain battery too badly. And doesn't eat up data if it's over wifi.

Yes this definitely happens. I've had people on the same network as me getting Japanese YouTube ads just because I or someone else looked at a few Japanese videos.

Load balancers call this IP Persistence. It's one way of keeping users connected to the application server that has their session in memory. Except in this case, the session is ad info about the user.

I think this is the only model for Whatsapp. They know who is talking to who. They don't know what you are talking about but that doesn't matter when one or two persons in a group show some interest in a subject.

If you or your friend visited the webpage of the product and it had a Facebook Like button then Facebook would know that you or your friend were interested.

If it was your friend that visited the product webpage, then Facebook also knows that you are friends with him and also knows that you are both interested in similar things based on your browsing history and so it would show you the ad.

What you describe is certainly technically possible, but could not be secretly done by a major corporation because it would easily be caught and become a PR and legal disaster

Welcome to the wonderful world of targeted advertising part #2. This and more is old hat in present day ad tech. How about tying together which user is the same person on a tablet, phone and a PC? And it doesn't need any kind of audio signal.

I had a weird experience some time ago. I stayed with my family at a hotel and our room had an unusually shaped tap in its bathroom. I never looked it up online - in fact, I didn't even mention to anyone that I thought that was a very peculiar tap. But a few days after we came back, Amazon showed me that exact tap as a shopping suggestion. It felt a little creepy. I could see no connection between the site where I booked the hotel room and Amazon, but there must be one somewhere.

Is it possible the manufacturer of the faucets actually partners with the hotel to market the faucets and they actually pay the hotel for the guest information to show them ads? I know some hotels I've stayed at actually advertise the fittings so it doesn't seem like a great stretch.

Still deeply creepy.

A previous or subsequent guest searched the tap online, and their search was associated with your account because it was from the same IP?

That actually makes a lot of sense. I think I used their wifi.

I'm very surprised this is news, especially on HN but even on the BBC. I can't believe serious IT security researchers wouldn't be aware of it.

A well-known (to me, I suppose) application is to listen to commercials playing on TV, in pursuit of the ad industry goal of tracking users between platforms. There's even a tech that ads an inaudible sound to commercials that your phone can pickup.

also related:

RSA key extraction in one hour via mobile phone microphone placed next to PC

http://m.cacm.acm.org/magazines/2016/6/202646-physical-key-e...

Not that long ago, I had an experience with this[1]. I'm currently running a test with my old phone to confirm that it is listening to me. I have an old android phone, with the microphone disabled (tape on the input)

Every once in a while, I will have the tape removed and say a few keywords, then disable it. My hope is to see if I will receive any advertising in with these niche keywords. It's only been a week so far, but I will update the post if there is anything.

(Note I have create a new google account just for it.)

[1]: http://idiallo.com/blog/be-mindful-with-ad-targeting (disclaimer: my blog)

As I've been suggesting increasingly on HN and elsewhere: sell me hardware without a mic.

I no longer trust mobile phones at all, so I carry one virtually never.

I use a tablet. It has an integrated mic and two cameras. The cameras are taped (they should have lens caps). On the rare occasions I use it for voice comms, I plug in a headset.

Pervasive audio pickup is a growing problem. It's going to be all but impossible to avoid other devices (other people's phones, audio-wired Internet-of-Shit lightbulbs, toasters, refrigerators, thermostats, video panels, Glade Stasi Vases (e.g., Google Now, Amazon Echo), cars, busses. But you can at least assert your own personal hardware.

On a related stance, Rob Gonggrijp and Frank Rieger of the Chaos Computer Club suggested a certification for "Guaranteed Cloud-less Objects", given the recent surge of IoT devices that spy on you, or lock you out unexpectedly because of server maintenance.

Source: Video at https://media.ccc.de/v/32c3-7501-ten_years_after_we_lost_the...

Interesting!

Thanks.

> Facebook also told the BBC it does not allow brands to target advertising based around microphone data and it never shares data with third parties without consent.

This and the follow-up paragraph read like:

> Facebook monitors background noise already without using it themselves for advertisement targetting purposes, but will share the data with third parties if you miss a privacy setting.

/tinfoil

Are you intentionally changing the scope of terms in your misparaphrase?

FB doesn't share data, it uses data for targeting, and it doesn't target by mic data.

If we accept that facebook is monitoring your microphone, and they are not targeting ads based on the mic data, then what is the point of having it monitor your microphone to begin with?

> It said Facebook ads are based only around information shared by members on the social network and their net surfing habits elsewhere.

so...cookies?

The fact that we have to ask that question and can't be sure of the answer is worth an article in itself.

It's being done because:

1) there is power/money to be gained 2) users will consent to it, actively or passively

#2 isn't really a requirement in emerging technologies and with little case law on the books

I'd be interested to see low-power voice recognition. My PowerShiri demo uses ~12% CPU on a Core2 Duo laptop when it's listening.

I'm pretty sure most apps use the default OS services, which just send the compressed voice data over to a server, where the actual recognition is performed (and your voice data stored forever, for "research and development" reasons...)

Such a stream would have been detected a long time ago. Do you have any evidence?

It would be trivial to have .NET make UDP calls based on speech recognition input. But what you're describing sounds more like a phone call, where all audio is transmitted, not just a few bits of data. If you have a pcap file showing the outgoing packets, I'd be interested to see it.

You need to try CMUSphinx http://cmusphinx.sourceforge.net. In keyword spotting mode it is very lightweight. Python example is here: https://github.com/cmusphinx/pocketsphinx/blob/master/swig/p...

It definitely does, and I wouldn't be surprised if big companies did not do it directly. I'd theorize that the tiny apps people download for games or little cat pictures unkowningly integrate this functionality as a dependency of advertising SDKs they use to make $0.10 CPC then the ad middlemen sell it back to the bigger companies.

I wonder how many (foreign) security agencies use exactly the same technology... so easy to access... so much to gain...

Every security agency is foreign to someone, and many have peering arrangements

Doesn't this need a visible permission in an app or are there workarounds like using the google play api?

You only need to put in the app an obscure functionality that makes use of the mic, and then explain what it is in the app description, for the 1% of people who will even bother to ask why Super Candy Smash Tournament is asking for permission to use the mic.

If Google were serious about granular permissions in Android it would allow you to choose from three options: "allow", "block", "pretend to allow" for every requested permission, where the third would be described as "Your app will be given permission but will not have actual access unless you grant it." Then later you could have a pop-up "This app is requesting access to your camera. You have previously chosen to give it fake access. (a black image.) this app is requesting camara access: allow access/block/pretend to allow?"

likewise contacts, photos, whatever. There should be detailed places to put fake content to give apps, for example here are my fake contacts, here are my fake photos, here are wav files for my fake microphone feed. Some people have new phones and haven't sent SMS's yet, or have no contacts - how would an app know if this is the case?

So fuck you, pinball gane app, the fuck you need to read my SMS and contacts and camera photos and need my microphone for. No, I just have a new phone and am in a quiet room :-D

Google's Android permissions model, application ecosystem, and vendor partnerships are utterly and irredeemably broken.

Not from the point of view of Google's business model.

That problem can be fixed.

Looks like xprivacy.

Didn't they add this to Marshmallow? So that if the app wants to access the camera roll, but you don't allow it, it just gets an empty list of images.

(I might confuse this with iOS, though. Does someone have a source?)

Unfortunately these OS level privacy countermeasures are trivial to circumvent.

Instead of checking for [photo permission denied] or [camera permission denied], on supported OS versions the app can check for [empty image list] or [black camera image]. Then display the impenetrable wall, or disable the goodie, or lock out the content because it can't verify that there are N or fewer viewers, or...

I addressed this explicitly: an advanced user could create fake contacts, take or upload fake photos (for example, I might not care about EXIF data), create fake SMS's. How should an app determine the difference between a user with a new phone with few or no SMS's, and a user who has chosen to fake a few or no SMS's? More to the point, the phone could simply generate fake content algorithmically, and these algorithms could change all the time. How would an app keep up? If it helps you think about it, at this point the phone is the bot and the App is the one trying to keep a bot from successfully using its API. Surely possible, but not "trivial".

I have a huawei $60 android which is 2 year old. I don't have a data plan nor a SIM in it, I just use it for offline openstreetmap data with OSMAND and for watching youtube video. I'm not so rich so I use a nokia candybar cell phone for my calls.

I was at this park, and my battery was nearly dead. At the 20% mark, I noticed that the battery drained very quickly, and it lost 1% per 10 second. Maybe it's a battery physics issue, or maybe it was emitting my position/recording to a GSM antenna... I guess it was defintely a battery physics issue.

Battery meters are nonlinear (i.e. they don't follow a straight line) in the regions close to full and close to empty. Battery level meters also tend to go out of calibration over time. I think these two factors together probably explain what you saw (as you say, a battery physics issue essentially).

Android 6 will allow to block the permissions of an app, right?

> Is your smartphone listening to you?

I wish. Not doing so would seem like a missed UX opportunity.

Reply on news.ycombinator.com