LSE - Small Logo
LSE - Small Logo

Charlie Beckett

March 31st, 2015

But how do you know that it’s true? Notes from #nishbr verification workshop


Estimated reading time: 5 minutes

Charlie Beckett

March 31st, 2015

But how do you know that it’s true? Notes from #nishbr verification workshop


Estimated reading time: 5 minutes

Notes from a Google-supported European Journalism Centre workshop in Hamburg on new ways of verifying user generated content for mainstream media organisations.


I wrote this in the London Observer last Sunday – I hope that the journalists who attend these kinds of workshops are making it come true:

“Newsrooms are investing in the new skills they need to carry out their changed role as curators, rather than controllers of information…we have to reinvent journalism once again. When aircraft plunge into mountainsides or when diseases ravage whole countries people still turn to news organisations, though they might find them through Facebook or Instagram.”

I also made a programme for the BBC recently about news. It showed that people don’t trust social media much but they don’t necessarily trust mainstream media either. The public  find much of hard news frightening, overwhelming and even if they are paying attention they are often confused.

This is not a happy environment for those of us who think journalism has a positive role to play in society. So what are you going to do?

Craig Silverman is right – you can’t compete with the bullsh*t, clickbait, aggregation merchants  – so for ‘real’ journalists good verification and curation is your business model not an add-on option.

By the way, I am delighted that Google are getting more engaged in this issue – showing that they are not just neutral digital intermediaries – but that they’re also trying to be a structure for ‘good’ information.

Some key points from the workshop – laced with a little of my thinking:

So first up Jenni Sargent  told us to change the terminology – forget UGC – it’s Eyewitness media

original photographs etc that are not planned and deemed to be valuable to publishers who distribute them through their channels –

For me that is networked journalism – journalists working with the participation of the public.

Interacting with this content should not just be about moments like disasters such as the Glasgow Crash. You can use more Eyewitness stuff to make news authentic not staged. Though of course, it’s with those sudden dramatic news events that having the public as accidental reporters is most useful.

Polis published a report in 2010 that showed how good journalists are at harvesting content from SM – as Michael Wegener from ARD described – but as Jenni Sargent said – we need to think about a proper relationship of verification that is not just making sure it’s true but also includes accreditation and citation of the source and the context of the subject.

Craig Silverman described journalism as a sense-making process designed to fill vacuum in anxiety-making situations.

He warned that you add credibility to rumour even if you repeat it – even with caveats (I’ll come back to that at the end).

Craig showed that the reluctace of the narrative to change – even when it is shown to be un-verified –  is caused by political frames as well as lack of verification. False reports of  people dieing is usually quite easy to disprove. The reports of a future period of solar darkness are by their nature harder to counter – especially as people may think it’s a joke and it didn’t make much difference to people’s lives. But false information about the Ebola scare does matter – yet it can spread because it confirms prejudices and creates panic.


There was a good example from Craig of the editing gap with the idea of Headline Dissonance. Our reluctance to correct the headline as well as updating the body of text shows we’ve lost the art of killing stories. Because there is infinite space online (but not infinite time to verify of course) we actually allow the overall impact of flawed information to remain even as we correct the details.

So we have to deal with it.

Craig’s advice reminds us that people are not always stupid but they do have very little time and are not paying attention. The public don’t want to have to do the journalism. Don’t be subtle. When debunking explain WHY people may not be telling the whole or any truth – it’s usually to do with profit or prejudice or propaganda.

Don’t just think as an individual journalist. Think as part of your brand. Is it obvious and are you protecting it? Be more overt about why you can be trusted and what you are doing with information to make it trustworthy.

I agree with Michael Wegener from ARD that all journalists have to master this. You can’t leave it to the “UGC verification team”. Outside orgs can help. The tools that people like SAM provide are excellent . Why waste time and resource re-inventing the verification wheel? But verification should be part of your all organisational production process.  It’s your identity. In that sense I am less bothered about exact common standards for all news orgs – you should create them to suit your own newsroom or market – but it is certainly in the interest of Public Service Media organisations or even public relations or NGOs to have public codes.

I liked Hauke Janssen from Der Spiegel’s point about the routine nature of verification. He reaffirmed how fact-checking has a long history and also how it can be seen to be counter to the journalist’s desire to be current, to be immediate, to be part of the flow – especially with breaking news.

And this leads to Hauke’s larger, almost philosophical-sounding point about how the rush to publish can do more than allow for individual mistakes – it can also confirm deeper biases and it can complete incomplete narratives. Here is a bigger warning: verification is only the first step to getting proportion and selection of narratives right as well as the detail. As Jochen Spangenberg from Deutsche Welle says it is also about factors such as Context.

And of course part of context is remembering that the internet is not real life and that twitter is representative of almost nothing.

Be Humble

Eliot from data investigation website Bellingcat showed that you don’t have to be a traditional MSM organisation to do this. MSM needs to be humble enough to connect to outside sources, but also sceptical of agendas that they might have.

Also important, as Eliot’s work shows, that you don’t get obsessed by just one platform.

And as presentations from Storyful and Google showed, it’s not just about the negative job of spotting fakes. The same tools can be a positive way to explore ‘hidden’ areas of public life such as hip hop chat.  Verification can be about creativity. This is about good storytelling – not just killing stories.

I think the most difficult part of this is how far do you go to show what is provisional. If you are not 100% sure can you still be ‘not wrong for long’? Craig is right that you give something attention even if you mention it to de-bunk it. The converse danger is that if an unverified narrative is out there you can look as if you are concealing or avoiding something if you ignore it entirely. With live broadcast – and we are all more like live broadcasters these days – you often have to have a conversation with the audience along the lines of ‘we too have seen that rumour too and we think xxxxx’.

Wait And Be Late?

If you wait, it can be too late. Journalism – especially around breaking news – has always been imperfect – we should admit our imperfections rather than retreat to a moral high ground remote from the public space.

And that brings me to my final point. Correct in the open.  I agree with Hauke that you gain more long-term credibility if you do so. Journalism is not a priesthood. And anyway,  we have seen what happens to priests who cover things up.

It was clear from the afternoon workshops that there are more and better tools around. We should never rely on algorithms but they are a big help. Ideally, they free us up to add editorial. They also free us up to be more critical about the data – as the brilliant work from Christina Elmer of Spiegel team on bad ‘cuddling’ research showed.

So verify the facts, explore the context, critique the data, converse with the audience, show your working and question everything. Most of all connect with your citizens and enjoy yourselves!



One of the best resources around on this is Craig Silverman’s Verification Handbook

Screen Shot 2015-03-31 at 12.20.25

About the author

Charlie Beckett

Posted In: Director's Commentary | Journalism


Comments are closed.