It's taken a while for me to get to like Twitter. As someone deeply concerned with the erosion of privacy and civil liberties, the self-surveillance implicit in constantly telling the world what you're doing was off-putting.

Well, I've got over that, and I'll be blogging soon about how I'm using Twitter for business. But it's occurred to me that I might not have been far off the mark the first time.

I'm using twollo.com to help build contacts with other people. It's a great tool. You provide keywords and phrases relevant to your interests and twollo.com tirelessly monitors Twitter conversations for matches. When it finds people tweeting using the words and phrases you've defined, it automatically adds you as a follower. I've found some very interesting people that way.

And I can see how it would be great for intelligence agencies, too. These organisations are skilled at recognising the words and phrases that might indicate activity of interest to them.

No, it's not as simple as watching for words like 'bomb'. They might, however, take a closer look at anyone mentioning the name of a radical imam, or a location which other intelligence sources suggest might be the focus of illegal activity. By itself, Twitter wouldn't yield much. But guided by these other intel sources, it could provide a means of discovering people not currently on the agencies' radar. And - through analysing followers, @replies and retweets - it could help map networks of people, regardless of the content of the messages.

With the amount of spying we're doing on ourselves - through Twitter, Facebook et al - we're practically doing the spooks' jobs for them.