That's something that we struggle with in the U.K., because we have an absolute prohibition in relation to surrendering to torturers. However, when we're looking at extradition, we're looking at something prospectively, so the test is whether there is a real risk of violation of article 3, which is the prohibition against torture.
The difficulty we now have is, I think, one that you were also grappling with. That is when that risk is now raised, if it is shown to be real and a diplomatic assurance can be given in response to it to say, “Whilst we recognize there's a real risk of torture, we guarantee that in this particular case, this person will not be tortured”.
I have quite a lot of difficulties with this, because the only reason you need a diplomatic assurance is if there is a real risk. You're then trusting [Technical difficulty—Editor] where you've already established that there is a real risk they're going to torture somebody to give you an assurance where they say, “But you can trust us [Technical difficulty—Editor] but we won't torture them”.
There are particular difficulties, as well, with monitoring. Very often, there is no monitoring mechanism built in, or the monitoring that a court suggests is going to happen is pretty fanciful. For example, they say, “Your client can complain if they are tortured”, and you say, “Well, I'm pretty sure they're not going to want to do that when they're in prison in this country, because they're obviously going to face further ill-treatment if they complain”. The methods suggested by the courts in the U.K. as sufficient to ensure that there will be no ill-treatment are, in my mind, not sufficient.
I think one of the things one has to grapple with when using [Technical difficulty—Editor] assurances is how you can make sure that they are going to be effective in practice and that they're going to be monitored, given that they're only given when there is already an agreed risk of this bad treatment happening.