Re: Crazy idea: How trust in darknets enables secure democratic censorship

On 7/12/05, Matthew Toseland <toad-EI5O+8PHWbJeeLb3ft/[email protected]> wrote:
> On Tue, Jul 12, 2005 at 12:49:37PM -0400, Evan Daniel wrote:
> > On 7/12/05, Matthew Toseland <toad-EI5O+8PHWbJeeLb3ft/[email protected]> wrote:
> >
> > > If the majority is wrong, it will disaffiliate from the minority. We are
> > > not talking about global voting here, we are talking about each node
> > > deciding on the basis of adjacent, trusted nodes. Yes there is some
> > > influence as far as majorities go, but the likely scenario is that the
> > > network splits into the two groups.
> >
> > Doesn't this mean that in the normal course of events, the majority
> > keeps the main contiguous network, while the minority ends up with
> > many tiny islands (potentially only 1-2 nodes) instead of an
> > "alternate" network?
> 
> Yes... sometimes. Hopefully the subnetworks would already be connected
> when they join up - look at the PGP/GPG web of trust model; you get
> fragments, and the fragments tend to connect with one another.

Maybe.  PGP et al seem to have easier ways of path folding available
than a darknet would -- I don't normally object to identifying myself
and verifying keys with someone I only know marginally, but I do to
exchanging freenet contact info, particularly when people's online
freenet personas are divorced from their node operating personas.  I
have no way to learn anything at all about the node operator from
within freenet, without risking my anonymity as a node operator.

> 
> > If that minority wishes to repair their network,
> > then they have to find each other through other, presumably more
> > vulnerable, channels, right?
> 
> Probably, if they didn't have any existing connections.
> 
> > It sounds to me like this has potential
> > to enforce groupthink through network value effects.
> 
> Please elaborate...

The central large network is against some content.  The fact that it
is larger makes it much more valuable as a network.  Therefore I am
inclined to act in such a way as to stay on the big network.  That
means participating in and reinforcing any existing groupthink, right?
 And that positive feedback loop means there will be groupthink and
that it will probably be strong and evident.

> >
> > Also, there is a chilling effect even without outing nodes, right?
> > Suppose I find freenet valuable as a source of semi-acceptable
> > content, but am also considering "marginally" acceptable content.
> > Wouldn't I be tempted to self-censor out of fear of losing my good
> > connectivity?  It seems to me that self-censorship to stay within
> > *percieved* limits is potentially even more limiting than those same
> > limits actually are.
> 
> That may be a point. There's no reason you can't be part of more than
> one darknet, but certainly nodes would self-censor to avoid having
> enforcement procedures taken against them.

When you say "part of more than one darknet," are you referring to
separate clusters within one large network, or entirely divorced
networks?

> >
> > And finally, let's suppose I use freenet for a variety of non-illegal
> > things, and don't partake in local votes to censor content at all (on
> > libertarian grounds).  In the event the govt seizes my node looking
> > for (nonexistent) illegal activities, aren't I likely to find myself
> > liable for failing to vote correctly, since the network has provided a
> > convenient method to do so?  And isn't this threat also likely to
> > create more negative voting than would otherwise occur, and thus
> > exacerbate the chilling effects?
> 
> If we don't provide such a mechanism, they will still find a way to make
> you liable for crimes conducted involving your node. The whole darknet
> architecture assumes nodes will eventually be illegal in and of
> themselves.

If I'm running an illegal freenet node, that means I'm willing to risk
the chance of being prosecuted for one count of whatever crime they
make it.  It doesn't mean I'm also willing to risk prosecution for 238
counts of knowingly allowing my computer to transmit child porn.  As
best I can tell, the law is reasonably tolerant of "I did everything I
could reasonably be expected to do" as a defense.  If the network is
such that I can't easily do anything about child porn, and there's no
evidence I personally looked at or posted it, then I'm at least
somewhat optimistic about avoiding prosecution.  If I actively failed
to vote against it through normal software features, then I'm a lot
less optimistic.

> >
> > I must say, I would be hesitant to use a censorship-enabled freenet.
> > Small communities with the ability to censor can be very oppressive,
> > and it seems entirely reasonable for me to want privacy from the same
> > friends I am connected to, even though I mostly trust them.
> 
> Well, you trust them enough to connect to them despite the fact that
> Freenet is illegal in itself (in the given scenario). Beyond that, your
> anonymity would only be blown (to them, not to others), if you posted
> something objectionable and it could be traced back...

I have friends who I would trust to not reveal that I'm running a
freenet node, even if it was highly illegal.  That doesn't mean I want
any chance of them discovering my personal taste in, say, fetish porn.
 Even if they would be willing to allow it on their nodes, I think my
privacy from them is important -- in some ways more important than my
privacy from the government, depending on exactly what content is
under discussion.

Evan