"Warrant canary" is a colloquial term for a regularly published statement that a service provider has not received legal process that it would be prohibited from saying it had received, such as a national security letter. Canarywatch tracks and documents these statements. This site lists warrant canaries we know about, tracks changes or disappearances of canaries, and allows submissions of canaries not listed on the site.
Follow us on Twitter for updates and notifications about canaries on this site.
Canarywatch is a coalition of organizations including the Electronic Frontier Foundation, the Berkman Center for Internet and Society, NYU's Technology Law & Policy Clinic, Freedom of the Press Foundation and the Calyx Institute. The Calyx Institute runs and hosts canarywatch.org.
For general questions please email email@example.com
Warrant canaries come in many forms. Service providers can adapt the design of warrant canaries to reflect the many different reasons why the providers want to speak about law-enforcement and national-security surveillance requests for customer information and communications.
While there is no one way to "warrant canary," every provider that wishes to publish a canary has to make choices about the following features:
Where is it published?
A warrant canary can be either a standalone statement or part of a more comprehensive transparency report. Standalone statements generally appear on a website, either as a separate page or as a simple statement published on the bottom of a website that says something like "Provider X has received 0 national security requests."
How often is it published?
Providers have to decide at what rate to publish warrant canaries. For example, some companies publish annual transparency reports with canaries. Some publish canaries bi-annually. Some providers simply leave a warrant-canary statement up until, ostensibly, it is no longer valid.
Apart from the frequency at which a canary is posted, a provider must consider whether or not to have a gap between the time period covered by a canary and the date the canary is posted—the so-called "latency period" of a canary). Some providers may issue a warrant canary that current through the posting date, while others may reference a particular, usually recent, time period. For example, a provider could choose to publish a report about the January—June period in August, meaning any canary contained in the report would have a two-month latency period.
A longer latency period, as well as reports that cover a longer period of time, give a provider time to ensure that they have done the due diligence required to put together an accurate transparency report. Additionally, a longer latency period could give the provider crucial time to challenge a gag order (although it is important to note that no provider has yet successfully challenged a gag order, and legal challenges can take years). Alternatively, shorter latencies might increase the possibility that a canary will change in a way that makes it appear a company has received legal process even if it hasn't. Of course, a shorter latency period means it becomes clear sooner to the public that a canary has "died."
What Kind of Request and/or Gag Order Is It Talking About?
There are many ways law enforcement can issue gag orders, including national-security letters, Foreign Intelligence Surveillance Court orders, court orders under the Electronic Communications Privacy Act, and others. Some providers issue warrant canaries for all national-security process, or for all law-enforcement gag orders. Others list the different legal authorities separately.
Legal processes that may be indicated by a warrant canary include:
How Do We Know It Is Real?
How a warrant canary is made public can affect the level of trust users have that it is authentic. Unless a system is in place to assure users about a canary's authenticity, users may question whether a warrant canary came from a provider or from someone pretending to be that provider. Some providers use encryption or other methods to verify the warrant canary as authentic, or to help expose a fake one. For example, Spideroak's warrant canary is "specific plain text signed with multiple GPG keys." (Of course, it's important to note that even if a canary is signed by a key, that doesn't mean that the provider wasn't forced to sign it.) Other providers simply publish their canary on a website run by the provider.