Hacker News Reader

Comment by swordswinger12 | Hacker News Reader

I have an honest, and probably dumb, question: how do warrant canaries actually avoid the prohibition against disclosing the receipt of a national security letter? Like, how is taking down a warrant canary legally dissimilar from just tweeting "we got an NSL"? If it semantically "means" the same thing, then why is one illegal and one legal?

View on HN

It's certainly a gray area, and a potentially dangerous one, so for legal protection if you take that warrant canary down then you would normally publicly state that it is absolutely not because you got an NSL, but for other reasons.

Like say, "business reasons".

I was looking into the limits of compelled speech during the Apple v FBI situation earlier this year. IANAL. The courts have said that there are some definite limits on when the government can force speech. I'm not sure where a warrant canary falls, but you should look at Wooley v. Maynard if you want to learn more[1].

[1] https://en.wikipedia.org/wiki/Wooley_v._Maynard

I think the idea is that a warrant canary is regularly updated. And the fact that it is NOT updated would be an indicator that a warrant (or whatever) has been served. So it's the INACTION that triggers it, not an actual action (like tweeting).

I suppose for that you would need to say "We have not received a letter as of $date" and then stop updating that date. The current notion, adding a line you'd remove if you did receive one, may be still an action, if it's part of your default reporting.

But I don't believe warrant canaries "actually work" in practice.

Yes, the apparent standard practice of just saying "We haven't received a letter" and then removing it if one is received makes no sense to me. Much better to have something you regularly update, and then stop updating it. The government could nab you for deleting something, but they might have trouble forcing you to update your declaration with false information.

Agreed. Although I wonder if removing cron or similar to update the date would be the same as removing the message legally speaking. I doubt these companies would have someone go in every day and bump the date. Then again that might be a good thing to outsource.

I don't think automating it would be a good idea. I'd say update it monthly or quarterly (so it's not too burdensome) and do it entirely by hand. Then when you receive a letter, just cease to do that.

I'd like to see it tried in court, anyone with two braincells can see that it signals that a NSL was received, something that's prohibited by an NSL. The intent and signaling are there.

I'm not a lawyer, but my understanding is that warrant canaries work through inaction. If a court were to require the continued publishing of the canary, they'd be crossing the line from something approximating a "temporary restriction", to compelled speech. And though I think neither are particularly defensible, the latter is much less so.

The key with warrant canaries is that they aren't removing a line from something previously published, but failing to publish an update to to it.

They're vague enough that it really doesn't reveal any useful and live in a weird area where preventing them from working would be compelling speech which courts have been iffy about. So while they're not providing any real information about the NSL other than 'at some point in the last $TIME_PERIOD we received a NSL (or Bob who pushed for the canary forgot to put the new canary up)' they're not enough of a break for the FBI et al to bother.

Reply on news.ycombinator.com