Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Australia’s vague anti-encryption law sets a dangerous new precedent

protonmail.com

Post reply on HN
Print view View original

221–230 of 265 posts

Re: Australia’s vague anti-encryption law sets a dangerous new precedent

#221

Post by cyphar »

cesarb wrote: post #59

For free software, I wonder if reproducible builds plus a "certificate transparency"-style check in the updater (only allow an update once several build servers, preferentially located in separate jurisdictions, have validated the build and published the corresponding source code) could help. That is, make it impossible to push a backdoor to a single user without making it public to everyone. Making updates anonymous…

I don't think that's sufficient. We need devices that only allow software to run that has been signed by TPM-resident keys on the device. Updates are only attempted to be installed if the binary has been signed by multiple keys (by people in different jurisdictions) and then the device prompts the user to sign the update (which requires entering the TPM passphrase). Even if you managed to compromise all of the developers you couldn't run signed code on the device -- you'd need the user to install it.

And for bonus brownie points we could have reproducible build checking (a-la certificate transparency) against the source repo, to see whether the binary is different to the official one. However, I think the threat model might have to be reconsidered (if all the developers are compromised, couldn't they upload a bad hash to the certificate transparency trail with a dummy version that only one user is given?).

I might write a blog post about this actually, though I'd need better experts than myself in IMA (which is what you'd use on Linux for this) and other secure-boot work.

Re: Australia’s vague anti-encryption law sets a dangerous new precedent

#222

Post by cyphar »

danieltillett wrote: post #178

I am an Australian software developer. There is no way I am putting any backdoor into any software I write and I am willing to go to jail if needed. If all us Aussie developers tell the government to go jump this stupid law will fail.

The punishment for non-compliance is civil fines, not gaol time. However I believe it's technically possible for them to push you into bankruptcy by making many requests and revoking them after you refuse them (fining for each copy of the request they resend to you).

Re: Australia’s vague anti-encryption law sets a dangerous new precedent

#223

Post by whiskers08xmt »

metta2uall wrote: post #218

How about we take a bigger-picture view than the implementation flaws of this super-rushed law and ask what is to be done about encrypted messages that allow many serious criminals to circumvent traditional police powers of search & surveillance? I think society as a whole will not accept criminals having such an advantage. So I think alternative laws have to be suggested & promoted, otherwise potentially really bad…

Coded communication is nothing new between criminals, and weakening the security/privacy of Australian software is not going to solve this problem. Weakening encryption of Australian software is not going to work against organized crime; These organizations are quick to adapt to changing law enforcement techniques. It is, however, going to make it much easier to surveil the general populace.

Re: Australia’s vague anti-encryption law sets a dangerous new precedent

#224

Post by metta2uall »

whiskers08xmt wrote: post #223
metta2uall wrote: post #218

How about we take a bigger-picture view than the implementation flaws of this super-rushed law and ask what is to be done about encrypted messages that allow many serious criminals to circumvent traditional police powers of search & surveillance? I think society as a whole will not accept criminals having such an advantage. So I think alternative laws have to be suggested & promoted, otherwise potentially really bad…

Coded communication is nothing new between criminals, and weakening the security/privacy of Australian software is not going to solve this problem. Weakening encryption of Australian software is not going to work against organized crime; These organizations are quick to adapt to changing law enforcement techniques. It is, however, going to make it much easier to surveil the general populace.

There will always be some criminals who are sophisticated enough to evade surveillance (until they make a mistake) - but that doesn't mean it's useless as many violent and/or organised criminals do get caught. Many get caught even with old-fashioned phone taps.

Also, aren't there ways of implementing targeted surveillance without weakening privacy/security very much, if at all? For example, targeting a specific user/device and making sure all exfiltrated data is encrypted with a public key belonging to the police.

Re: Australia’s vague anti-encryption law sets a dangerous new precedent

#225

Post by metta2uall »

Taniwha wrote: post #129

(essentially repeating a recent twitter thread here) Imagine you work in a modern software house and you get one of these ... and here I mean you, not your boss, not your coworkers, the govt knocks on your door and demands you put a back door in the thing you are working on at work ... So you write the code ... how do you write the unit test? how do you get it past the code review? the mandatory QA tests? ... all the…

I expect they won't ask for anything crazy like that or violate the prohibition on asking for a "systemic weakness" - one of the reasons being that they want this law to be deployed as smoothly as possible to make a good case study for their 5-Eyes mates.

Re: Australia’s vague anti-encryption law sets a dangerous new precedent

#226

Post by cyphar »

askvictor wrote: post #207

Consider Signal, which is open source and not based in Australia. If AU wants to intercept a signal message, then presumably they would need to either force Google and/or Apple to push a custom app to a specific user, or take over the entire phone (again, via Google or Apple). In the first case, is the app that comes from the app store somehow verifiable, or do you need to build from source to be sure? Is there anyth…

I think that, for long-term security, we need to have devices that are resilient to orchestrated sabotage by the vendor. The current approach by Apple is great, until Apple is compromised in one way or another.

I have some idea for how this could be done (TPM-resident signing keys on each device, which have to sign all binaries before they can execute). I might end up writing a blog post about the idea.

Re: Australia’s vague anti-encryption law sets a dangerous new precedent

#227

Post by gruez »

cyphar wrote: post #226
askvictor wrote: post #207

Consider Signal, which is open source and not based in Australia. If AU wants to intercept a signal message, then presumably they would need to either force Google and/or Apple to push a custom app to a specific user, or take over the entire phone (again, via Google or Apple). In the first case, is the app that comes from the app store somehow verifiable, or do you need to build from source to be sure? Is there anyth…

I think that, for long-term security, we need to have devices that are resilient to orchestrated sabotage by the vendor. The current approach by Apple is great, until Apple is compromised in one way or another. I have some idea for how this could be done (TPM-resident signing keys on each device, which have to sign all binaries before they can execute). I might end up writing a blog post about the idea.

Sounds like applocker in windows

Re: Australia’s vague anti-encryption law sets a dangerous new precedent

#228

Post by jsjohnst »

Taniwha wrote: post #129

(essentially repeating a recent twitter thread here) Imagine you work in a modern software house and you get one of these ... and here I mean you, not your boss, not your coworkers, the govt knocks on your door and demands you put a back door in the thing you are working on at work ... So you write the code ... how do you write the unit test? how do you get it past the code review? the mandatory QA tests? ... all the…

> So how do I find out which software on Android Play is written in Oz?

If the above bothers you, then you should avoid any software written in the US too. The exact scenario you described can happen in the US, especially if you have a security clearance, but work in a regular tech job.

Re: Australia’s vague anti-encryption law sets a dangerous new precedent

#230

Post by gtbc »

cyphar wrote: post #157
gtbc wrote: post #149

Earlier quoted context omitted.

Does that have much to do with the matter at hand?

"The jurisdiction of the court is largely granted by statute" is not an entirely accurate statement (though there are restrictions on what you can sue the Commonwealth for) . That was my point.

You were referring to state laws. Review of decisions resulting from laws passed by parliament is granted by the ADJR act, and hence by statute.
Post reply on HN