Nieman Foundation at Harvard
HOME
          
LATEST STORY
With elections looming worldwide, here’s how to identify and investigate AI audio deepfakes
ABOUT                    SUBSCRIBE
April 24, 2012, 9:06 a.m.
LINK: blog.storyful.com  ➚   |   Posted by: Joshua Benton   |   April 24, 2012

Malachy Browne of Storyful — “we deploy the ‘Human Algorithm’ to deliver international stories told through the social web” — has a good blog post detailing how his outfit works to verify (or debunk) videos, photos, and text from social media platforms:

Within an hour of the Pacific tsunami alert being issued on April 11, 2012, Twitter and YouTube abounded with videos purporting to show monster waves striking the coast of Sumatra and Aceh in Indonesia. However, these were versions of the devastating 2004 tsunami and other dramatic videos, reissued with the 2012 date. In December 2011, when a police officer was killed at Virginia Tech in the US, a picture of the 2007 massacre was widely circulated as the 2011 event. In both instances, Storyful was quickly able to debunk this content as false.

But how to verify what’s true? The challenge for news rooms in this new age of information overload is who and what to trust.

A lot of it is good old-fashioned journalistic skepticism, but the basic m.o. — search the media object for checkable details, use available digital tools to check them — is good advice.

Show tags
 
Join the 60,000 who get the freshest future-of-journalism news in our daily email.
With elections looming worldwide, here’s how to identify and investigate AI audio deepfakes
“They are easier and cheaper to create than deepfake videos, and there are fewer contextual clues to detect with the naked eye.”
Google tests removing the News tab from search results
The News filter disappearing from Google search results for some users this week won’t help publishers sleep any easier.
Wealthier, urban Americans have access to more local news
But roughly half of U.S. counties have only one news outlet or less.