It's also impossible. "I'm going to kill some soldiers when I get back" might refer to my super secret terror plot, or my PlayStation-related plans for the evening.

Why not? We expect and demand that banks monitor billions of transactions between 100s of millions of customers for evidence of money laundering and terrorism financing - and fine them billions of dollars when their controls or oversight are deemed insufficient.

I expect most us agree that banks _should_ be held accountable for the legitimacy of their customers behaviour. Why should the same standards not be demanded of internet companies?

Is this just the first offer in a bid to get police access to data for investigation purposes - like ask for the world and then asking just for a mountain sounds reasonable.

In this case it would be asking that service providers monitor everyone; but, the actual desire is to allow investigative authorities (police, MI5, ...) access to suspected terrorists online dealings via the service providers at the back end.

In the Lee Rigby case it seems that the exchange about murdering soldier would have been sufficient to push the suspect in to the camp of making terrorist threats and allowed for an arrest.

It's curious in some ways as Adebowale had been the focus of investigations and yet GCHQ clearly didn't have information on all his online exchanges (that could be tied to him), yet a company in the USA did have access. This suggests that the level of integration in to online communications that GCHQ has is far less than we've recently been led to believe. We're being told left-and-right that government spooks know our every move and whisper - this seems not to be true, not even for people associated with terrorism investigations.

I don't think it's practical to require the likes of Facebook to monitor for terrorists and the like. After all one man's terrorist is another's freedom fighter. However it would be quite reasonable for them to choose to look out for that and bullying, grooming and the like especially if the posts are semi public.

I thought the global intelligence services already had access to Facebook etc. data, maybe the question in-light of the Ed Snowden leaks, is why they didn't see this. Except, as commented there would be an extreme false-positive rate. In which case, why do we continue to fund the intelligence services...

The title of this article is slightly misleading. The article actually says we should be monitoring known terrorists. Perhaps they should change the title to "Internet companies should not be monitoring non-terrorists".

I think that's a bit hard to do when your business model is built around user profiling, even if it is only for purely legal business ends.