Limits to Twitter as a News system

There is (to my mind anyway) a rather disingenuous post up on Techmeme today on the use of Twitter as a news service, based on its usage over the Mumbai attacks:

While CNN and other mainstream publications have been slow to credit online services like Twitter in the past for breaking news during times of crisis like earthquakes and wildfires, now they’re definitely coming around. It shouldn’t really be a surprise the CNN is the one saying this either, one of their anchors, Rick Sanchez, has been very active in the Twitter community, using the micro-messaging service to help with his reporting for several months now.

“Yes, Twitter is a source of journalism,” as blogger Mathew Ingram puts it.

But CNN isn’t quite as rosy about Twitter’s “reporting” capabilities as Ingram — after all, they don’t want it taking their jobs. Noting the widespread rumors and inaccuracies that began to pop up on Twitter about the attacks alongside the true reports, CNN warns that “while Twitter remains a useful tool for mobilizing efforts and gaining eyewitness accounts during a disaster, the sourcing of most of the news cannot be trusted.”

That, of course is true, but as Ingram notes, when some wrong information started appearing on Twitter about the attacks, other users were quick to correct the reports — and again, those corrections probably happened quicker than they would have through traditional media.

It’s really a question of speed versus 100 percent accuracy. In many cases you have to take one over the other. In a time of crisis, I would guess that just about everyone would choose speed. And in that regard, outlets like CNN simply cannot match a service like Twitter which has millions of potential “reporters.”

The old adage, “don’t believe everything you read,” is still important to keep in mind when getting information from Twitter. But it’s precisely an “old adage” because it’s been around long before Twitter ever existed and will be around after it’s gone. Twitter is just another source, a very fast one, but it still requires the ultimate filter: Your brain.

I didn't see it that way at all - I think the point Tweetip (who aggregate Twitter signals around a subject ) made about the pattern resonated with what I saw (see the graph above):

- First there is a signal [news], then noise, then the bandwagon.

And in terms of ratios I'd estimate it followed the Social Network standard of 1 : 9 : 90 in terms of percentages - ie c 1% was signal, 9% was noise around the signal. The "noise" I'd classify as the standard stuff you'd see around a terrible event - are people OK, is my friend OK, thats sort of thing - but the bulk, the 90%, was pure cr*p, with stuff such as:

- Wild surmises to gain attention
- Re-tweeting mainstream media in an endless circle jerk
- Mediawhores jumping on the bandwagon with crocodile tears and faux sympathy
- Hatemail
- Hatemail to the hatemailers
- People arguing with any or all of the above

What this means is that this was not news - it was a hose of sewage, in which the few nuggets of real news were virtually impossible to pick out. And the "corrections" were meaningless, as they were just adding to the morass of manure as it all swirled around in unverified retweets. Was the original right? Was the correction right - uh oh, here comes an opinion that the correction was wrong.

The old adage that you shouldn't believe everything you read was built for the days when what you read was at least curated and edited in some form, and you could apply rational judgement to it. When faced with a sewer pipe of cr*p of every imaginable sort, its impossible for anybody's brain to filter out what is "real" news and what is not.

If Twitter's performance on this is the "future of journalism", I don't think CNN has anything to worry about just yet.

What will be needed is ways to filter this - I can see 3 approaches that can be used:

(i) Verification - some way of stopping 3rd parties entering and hijacking the stream. with different agendas
(ii) Increased Transaction costs, ie some form of "strong tell" - only people fairly strongly involved will want to communicate
(iii) Filtering - probably combined algorithmic / empirical - to strip out clear recirculatory or off topic responses. Over the last 2 days I have lost any faith that pure "social media" can self correct fast enough or accurately enough in these situations.

How to do this is a fascinating task in itself, and when it starts to occur then I think we will see real "Citizen Journalism" occur, and at that point CNN can be worried.

Or maybe not - yesterday at Amplified 08, one of the sessions led by the BBC team was on the interaction of "Citizen Journalism" and Mainstream media - and I came away with the thought that at the scales the MSM operate, half baked shamateur systems are just not going to cut it. It will need big iron and serous organisation. The MSM have that. So, of course, do others, but it is MSM's to lose.

Update - not just me who saw it this way:

Update - as a friend noted, the CNN piece was rather a good analysis of Twitter, whereas the reverse was not true

Update - a day later, other people are starting to think about how to make Twitter more useful