[AusNOG] Assistance and Access Bill moves to PJCIS

[Robert Hudson]

On Tue, 27 Nov 2018 at 16:04, Mark Newton <newton at atdot.dotat.org> wrote:

>
> On Nov 23, 2018, at 4:46 PM, Robert Hudson <hudrob at gmail.com> wrote:
>
>
> On Fri, 23 Nov 2018 at 14:47, Paul Brooks <pbrooks-ausnog at layer10.com.au>
> wrote:
>
>> In theory no - this bill doesn't weaken encryption, and explicitly
>> doesn't allow any
>> changes that would weaken encryption.
>>
>
> They say that - but I don't believe them.  I don't think they even
> understand what they're suggesting (or if they do understand, they're
> relying on others not understanding, or not caring).
>
>
> I think it’s dangerous to assume they don’t know what they’re asking for.
>

To clarify - I was speaking of the politicians.

>
> MPs probably don’t know, that’s true. But they aren’t the source of these
> Bills: No has ever climbed out of bed in the morning and thought, “Y’know
> what ASD needs? Unencrypted access to SnapChat. Let’s make it happen.”
>

I agree entirely.

>
> MPs also aren’t in charge. PJCIS reliably decides whatever the bloody-hell
> ASIO and ASD want them to decide. The belief that there are a bunch of
> level-headed independent-minded politicians *making decisions* is crazy,
> there’s never been any evidence that that’s true.
>

I think you may have missed highlighting the ludicrous notion of *level-headed
independent-minded politicians*.  I'd put a smiley there, but the current
state of our political leadership (if one could call it that) is so abysmal
that it's no laughing matter.

>
> These Bills are drafted by the intelligence agencies themselves, and they
> know precisely what they’re demanding, they know precisely what the flow-on
> effects will be, and they’ve judged that for their own purposes, the
> cost/benefit analysis works in their favor.
>

This is the bit that I don't get.

They *must* know the effective outcomes of the TAN/TCN/TAR activities is to
introduce systemic weakness in the encryption processes they touch.  The
attack vectors against encryption (be it data at rest or data in flight)
are so narrow (given that they're asking for this, we can, I believe,
safely assume that they're not able to brute force things at this stage) as
to effectively mean "a way to retrieve the keys" or "a back door" - both
processes, once established, immediately introduce exactly the kind of
weaknesses the proposed bill supposedly protects against (noting the
incredibly low standard of proof that needs to be produced here).

And even when they manage to convince Apple, Google, Samsung, etc to hand
over unlock keys to phones, and convince Facebook et al to either introduce
back doors or back-channels into their messaging apps (they must know the
folly of asking a carrier to do anything with an encrypted bit-stream -
maybe the focus on carriers is to try to get them to inject unlock code
into the bloatware they load on phones), they *must* know that they simply
won't magically gain access to communications between criminals (by
whatever measure you define criminal, be it terrorist, paedophile,
organised crime, etc - anyone who is rightfully the focus of legitimate
law-enforcement activity) because any of them with the ability to tie their
own shoes will immediately switch to communications processes and systems
that are not subject to this bill.

The net result of this bill, like previous thought bubbles as the Internet
paedophile filter ("oh noes, Australians can't consume child porn any more,
oh well, we'd best wind up our little industry now, without the tiny market
that is Australia, we're clearly no longer viable"), will be to send the
real criminals, the ones smart enough to do real damage, deeper into the
places they're hard to find - they will just be driven further underground,
with no material impact on their ability to carry out their goals.

So, what benefit to the intelligence agencies get?  The power to track
terrorists not capable of finding the safety switch on an AK-47?  We seem
to be able to do that already, so I'm not sure that's something we can
accuse them of wanting.  Do they want to spy on law-abiding citizens (which
is contrary to the scope of their operational focus for some of them) - Is
this their real target?


> The possibility that the cost/benefit analysis works against other people
> is also well understood, but they choose to not distract the argument by
> engaging on that point. Bring it up as much as you like, they just ignore
> it and talk past it.
>
> For the last decade, there have been arguments about this stuff that have
> been based on the belief that the Government is too dumb to know what it’s
> asking for, and that reason will prevail if we just explain it to them with
> the facts.
>
> In case nobody’s noticed, that approach hasn’t worked, and there’s no
> indication that it will ever work.
>

I only carry this point because I believe it helps to highlight what the
REAL desired end-state may be.  Because of the technical detail, this won't
help to catch competent criminals.  It won't help to catch incompetent ones
either (because they largely already give themselves up through stupidity
and shithouse OpSec).  So who is left as the target?

>
> This community has spent years wasting its time by communicating facts to
> them that they already know, and don’t care about.
>

I still don't think the politicians really get it - but I do take the point
that faced with taking advice from the departments they preside over, or
the public and/or industry associations, when there's simply no negative to
ignoring the latter groups, means that we're not going to get listened to.

>
> They also don’t care about compromises: If you give them 50% of what they
> want, they’ll come back 18 months later and demand the other 50%. That’s
> how they’ve always worked (cf: data retention: The AA Bill is the grab bag
> of stuff the A-G couldn’t ask for last time. And if they don’t get it all
> this time, they’ll be back in 2021 for the next tranche)
>

I totally agree with this.  What the agencies don't get now, they'll simply
play the long game and get later.


>
> Victory on these matters will never be won by having an argument based on
> the assumption that they need experts to explain facts and technology to
> them. The only way victory will be achieved is politically: There needs to
> be blowback, asking for more will need to cause them pain before they’ll
> stop.
>

So, this needs to become an election issue - it's the only thing the
politicians understand.  We either need to convince the opposition or the
(ever growing) cross-bench that not only will supporting this legislation
lead to them not receiving votes in the next election, or that, more
specifically, opposing it will result in more vaults (offer the carrot,
rather than the stick?).  And make them realise that changing their mind
later will result in us changing our minds.

Or we form a political party (or we directly infiltrate an existing one)
and push a very specific agenda against this sort of thing.

By all accounts, we have until May 2019.


>   - mark
>
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.ausnog.net/pipermail/ausnog/attachments/20181128/3b62a155/attachment-0001.html>