Hacker News Reader

Comment by janvidar | Hacker News Reader

Isn't this really just a sign of flawed hardware design?

In my opinion hardware should be designed so that the camera LED lamp should always be lit if the camera is used. If there is a malfunction with the LED, then the camera should also not work. Also there should be a hardware LED for when the microphone is being used which should work in the same fashion for laptops with built-in microphones.

In the webcam drivers I have looked at the LED is controlled independently of capturing, although drivers do enable the LED when the camera is used. This essentially means that hackers can record and disable the lamp.

I've been considering hacking together some piece of software that will continuously use the camera (/dev/video) in order to block it for other applications, and have it fail with visible alerts if unable to block the camera. Not sure if the same thing can be achieved for the audio recording devices due to multiplexing.

View on HN

>If there is a malfunction with the LED, then the camera should also not work.

Many would argue that this is the more flawed design.

Reminds me of this: https://9to5mac.com/2016/02/05/error-53-iphone-6

It seems like a "fail safe" to me. The current design is a bit closer to a "fail deadly" in that it creates a mode that's the worst-case from the user's perspective: the camera works but the indicator doesn't.

It is probably worse to have an unreliable indicator light than it is to not have any indicator light at all.

In my opinion hardware should be designed so that the camera LED lamp should always be lit if the camera is used.

That would only inform the user that the camera is on, and that isn't good enough. If the user is paying active attention to the computer then they can take action, but if they're not then they can't - noticing an LED has switched on when you're busy doing something else in the same room is quite unlikely. Even if you're actively using the computer you could miss it if you're focused on a task.

A hardware switch that disables the camera or a lens cover that blocks it entirely (eg tape) are the only safe options.

The hardware switch won't tell you that your computer has been hacked. The camera light will. And a lens cover won't protect your microphone or your personal data.

If you are hacked, the camera light won't tell you. The camera light also won't protect your microphone or your personal data.

But the light won't tell you because of a flawed design. It should be trivial to design a webcam LED which can not be turned off while the camera is being used.

It should tell you. That's the whole point.

If I'm hacked and the attacker doesn't touch my camera, then it can't tell me.

That's the whole point.

But if they do touch your camera then you will know. If they don't touch your camera, then putting a piece of tape over it is pointless anyway.

The issue is the camera being used to violate your privacy. A hardware switch that lets you physically disable the camera, or a a lens cover that stops it seeing anything, stops the possibility of that happening completely. You'd obviously need a second system in order to stop an attacker using your microphone.

All a light does is inform you that the camera is active. That is not a defence. It doesn't tell you that you've been hacked because an attacker might not access the camera (a mic and data are still accessible), and the camera light might activate if you haven't been hacked (eg any piece of software that you're using might result in the camera activating regardless of whether you want it to or not) which would lead to many false positives and as a consequence people would ignore the light.

A light is only information, and any amount of information won't stop an attack. Information is only available after you've been attacked, so you haven't stopped or mitigated the problem. A hardware solution that stops people accessing the camera mitigates the result of an attack before it happens. That's why it's much better. You just have to remember that blocking the camera only stops attacks that use the camera. You still need to be diligent about other attack surfaces.

I really don't understand this reasoning at all.

If you just block the camera you won't know that your computer is being hacked. The piece of tape isn't going to stop the hacker from doing much. Knowing my computer is hacked is far more important than stopping the attackers from seeing me.

If you have a false positive that is also very concerning and should be dealt with. At the very least I can put a piece of tape over the camera after the light turns on, if I believe it's a false positive.

At the very least I can put a piece of tape over the camera after the light turns on

At which point it's already too late - the hacker has had access to your camera for some period of time, until you notice the light. If you're interested in stopping people seeing you then you've failed. Putting a hardware block on the camera (eg tape) before you're hacked means that the hacker won't ever get access to your camera to see you.

The hardware block is about stopping people accessing your camera. It has no effect on anything else.

Regardless of what you do with your camera you still have to defend against other ways you can be hacked.

>At which point it's already too late - the hacker has had access to your camera for some period of time, until you notice the light.

Yeah for like 10 seconds. Who cares? The problem with webcams is they can spy on you for days and weeks on end, and record your sensitive conversations.

Recording conversations has nothing to do with the camera. You might be thinking of the microphone. Should we put a little LED on that too?

Yes of course. It should be connected to both. You can't put a piece of tape over a microphone.

But the microphone is always on if you enable voice activation, e.g., "Okay Google", "Hey Siri", "Hey Cortana". That recognition is handled in software, so no "hacker proof" hardware indicator can distinguish between the use cases.

"Hey Siri" works on desktop devices: https://9to5mac.com/2016/07/15/how-to-enable-hands-free-hey-...

"Okay Google" was removed from desktop: http://www.theverge.com/2015/10/16/9553051/ok-google-removed...

Yes, and I have my own issues with unprompted voice activation for that exact reason.

I think that having the mic light always on when you enable voice activation, would go a long way to showing consumers the risks of it. It makes it visually obvious that you are being recorded.

Then a camera light that tells you when the camera is on, and a hardware switch that prevents image data from going to the computer.

Take it a step further - an auto-off hardware switch.

When you go to Skype you have to flip it before audio/video works - and then once the webcam is no longer "in use" it flips the hardware switch so that it can't be triggered (as the user is not going to want to re-trigger it, nor is likely to remember to do so)

Here's the problem with this solution - it assumes that everyone equates the LED being on all day long every day with "oh shit, I'm hacked". No, the average person is going to just say "Oh great, another thing's broken on my computer."

Your average computer has a lot of LEDs active anytime it's plugged into the wall (let alone running) - what protection will yet another LED offer to your average consumer?

Reply on news.ycombinator.com