Social media offers new types of information and intelligence. There are teams of cyber vigilantes that screen-grab extremist material online, save it, and report it. It has been an extremely effective way for citizens to keep watch on certain groups, and even pass hateful material to the authorities. Many of them do great work.

As I argued in #intelligence, a paper about social media intelligence, drawing on these citizen vigilantes will be a valuable resource in future. However, this also brings serious difficulties: misinformation is easy to manufacture and spread. Trolling is common, accusations and counters fly daily.

On Friday a screen grab of a Tweet allegedly coming from the English Defence League's official account was widely circulated. It read, in reference to the Newtown shootings, "If those kids were children to filthy leftists, Adam Lanza did the right thing This world is being destroyed by leftists #Newtown #EDL #NFSE".

Horrendous, and on seeing this I offered a note of caution about its veracity, suggesting it could be fake (for which I was accused of being an EDL apologist). But, how to tell?

First, motive. The EDL has a habit of saying very offensive things - members have been known to say supportive things about Anders Breivik. But justifying the completely random, non-political, massacre of children from their official account seemed to me a new departure. That doesn't mean they wouldn't of course.

So second, evidence. Although re-tweeted hundred of times, the evidence appeared to be a single saved screen shot from one individual, who said it had appeared only momentarily before being deleted. This is unusual given thousands of people closely monitor the EDL account. But as proof, this individual also posted a video, which showed the still open page of the tweet from where it was grabbed. The URL of the offending Tweet was 241922092158143412. This is not just a random number: Tweet URLs are sequentially time stamped. A stamp starting 241 - dates the Tweet to some time in September. Friday it was clocking around 279 -. It appears prima facie to have been doctored, using an older Tweet. Creating a fake Tweet (by editing the rendered HTML or even photoshop) is very simple indeed. Moreover, even deleted tweets are still accessible when searching Twitter's Application Programming Interface, which is then made available through third parties (although they should respect deletions, they often slip through). I'm not able to find any it anywhere.

This neatly illustrates the minefield of using what we termed 'SOCMINT' (social media intelligence). Online information is extremely useful, but scepticism is always required: especially with screen-grabs. Although many angrily re-tweeted it and even tried to alert new outlets, there were plenty of others ready to investigate: both an anti-EDL group and programmer came to the same conclusion as me independently. Perhaps the EDL admins are more technically skilled than I give them credit for, and the URL date discrepancy can be explained. In that case, I'll correct this article as soon as I can. I'm not interested in defending the EDL, but little good comes from spreading falsehoods. The EDL, in my experience, say enough offensive things of their own accord: there is little need - or excuse - to make any up.

Follow Jamie Bartlett on Twitter: www.twitter.com/JamieJBartlett