Talking reliable social sources…

This past Sunday, I went on CNN’s Reliable Sources to talk about the transformative power of crowd-sourced news – and how traditional media companies can best leverage the eyewitness accounts, expertise and varying perspectives of regular folk to improve the breadth, depth, and relevance of their coverage.

During my on-air conversation with Wall Street Journal tech columnist Walt Mossberg I was struck by his reluctance to accept the notion that innovative new technologies may soon provide the answer to what he rightly identifies as a pressing need to assure the reliability and trustworthiness of social sources and posts.

I of all people understand. It seems like eons ago now, but I started by professional life as a business and media correspondent for the Times of London. In my early twenties I was breaking big exclusive stories in the old-fashioned ‘shoe leather’ way. There was no Internet. Not even one of those ‘brick’ phones. My day was all about talking to and qualifying as many sources as I could – establishing their expertise, trying to figure out if they had hidden agendas, putting their words in context of both my cumulative knowledge and what other sources were telling me.

It was hard work, hampered by the limitations of the clock: Not enough hours in the day to talk to as many diverse sources as possible about the stories I was working on let alone the stories I never discovered or wrote up.

Now, as the founder of a technology company endeavoring to replicate with algorithms and automated processes as much of that investigative process as possible to parse social posts for accuracy, I think back to the ‘good old days’ and think how cool it would have been to be able to use the power of Twitter in real time – to find and qualify sources, discover new perspectives, get live eye-witness accounts. How much better my stories would have been if I could have put the ‘two source rule’ on steroids and been able to have interacted with hundreds of sources, each with one small piece of a more textured, in-depth and informative puzzle.

Yet when most journalists hear about crowd-sourcing news they immediately put up their defenses.

Sure, nothing replaces a reporters’ gut instinct, the ability to look someone in the eye and question them to figure out if they are lying (or conveniently omitting key details). Algorithms cannot yet replicate a reporter’s ability to detect patterns over time to put information in context but companies like Verifeed working towards achieving just that.

Crowd-sourcing the news is not going to eliminate any good print journalists, producers or TV correspondents. Instead using a platform like Verifeed and others like our friends at Cont3nt.com, Storyful can help improve the quality, range, diversity and speed of their output.

Reporters like Walt Mossberg whom I spoke with on CNN see value in technologies that can unearth live video of plane crashes, natural disasters and other breaking news stories. But it’s hard for them to imagine what would be credible enough to use in a professional reporting piece. Yes, irrelevant, false, and inaccurate posts are a problem. Remember that hacked AP Tweet in the spring about a “bomb” in the White House? It sent the Dow plummeting 128 points.

Could Verifeed have caught that? Well, we sure would have flagged it for closer inspection. Why? It was only ONE source. Can you imagine the volume of activity on Twitter if a bomb truly had exploded in the White House? Algorithms can look for patterns, cross-reference posts with source expertise and much more.

Think back to the Boston Marathon Bombing. When the FBI requested the help of the public in identifying the suspects, an interesting phenomenon occurred: Two blog sites 4chan and Reddit set up forums where people could post information they believed could be evidence. Participation on the threads took off. Hundreds of comments, photos and videos were posted. As can be imaged, a lot of the information was of poor quality. But Reddit had a major advantage: the ability of the crowd to “upvote” and “downvote” information based on its relevance and accuracy. Real time peer review.

Imagine easy-to-use processes that allow you to filter and parse all social information in real time for relevance, weeding out all the repetition and unreliable posts. What if the algorithms could search automatically so you can discover first a new trend, breaking news development or market-moving report?

Verifeed has architected and continues to advance these algorithms right now.

A program that automatically analyzes follower ratios, interactions, tweet frequencies can help, but even a high ‘influencer’ score on Klout and lots of RTs does not mean the Tweet is accurate. Lots of people re-tweeted the erroneous hacked AP Tweet.

What about software that can immediately eliminate untrustworthy accounts and give weight to valid accounts? Instead of pushing your way through a massive crowd trying to catch tidbits of valuable conversation, you’re suddenly sitting in a quiet restaurant where you can sit and dine on the finest information.

But even trusted sources can get it wildly wrong. Back in my BBC TV News anchor days, we erroneously reported the death of Bob Hope. Whoops. And in the hours after the Oklahoma City bombing, CNN and other sources erroneously reported the destruction was at the hand of Muslim terrorists. There are too many examples to recount, and any journalist, even an award-winning one, knows they’ve made mistakes.

Let’s go back to the crowd. If someone yells, “I’ve got big news!” you might think “OK, I’m sure other people are talking about this, I just can’t hear it right now because of all the noise.” But if you’re in a quiet restaurant you can hear all the conversations. If only one or two people are yelling, you may think, “If this news is so big, why isn’t anyone else talking about it?” An algorithm trained to pay attention to frequency of content as well as validity of sources will not be fooled by a single source, no matter who they are.

Of course, the story can only be as good as the available information – and insightful accomplished reporters who know what questions to ask.

It has always been the question that advances the story, that helps put developments, perspectives, assertions and statistics in context, and qualifies the expertise and credibility of the source.

That’s why we will always need journalists to be tenacious interviewers and questioners.

And while crowd-sourcing algorithms can mine social data that tells you the “when, what, why, where, how” and sometimes even the “why”, technology cannot write elegant prose, create compelling video or engaging story-telling.

As an award-winning print, radio, TV and digital journalist, I understand the challenges of verifying information and putting it context to be both first and right. At Verifeed we want to put verification tools at the fingertips of journalists so they can best leverage the energy and expertise of the crowd to get important untold stories told.