Nieman Foundation at Harvard
HOME
          
LATEST STORY
Deepfake detection improves when using algorithms that are more aware of demographic diversity
ABOUT                    SUBSCRIBE
May 14, 2012, 10:03 a.m.
Audience & Social
Truth Goggles sample

How to peek through Dan Schultz’s Truth Goggles, the B.S. detection software, right now

An MIT grad student is studying users’ behavior to understand how his software might influence their critical thinking skills.

Truth Goggles sample

Last November I wrote about Dan Schultz and his Truth Goggles, a piece of software he’s developing in partnership with PolitiFact that highlights suspicious claims in news articles. It’s as simple as spellcheck.

Truth Goggles logo

Now, as Schultz finalizes his graduate thesis at the MIT Media Lab, he’s conducting a user study to understand how people (anonymously) respond to a prototype of the tool. It’s our first look at the software in action, and he wants you to give it a try.

Here’s how it works. First you’re asked to evaluate the veracity of several statements. “Over 40 percent of children born in America are born out of wedlock.” “Some billionaires have a tax rate as low as 1 percent.” No Googling; you have 20 seconds for each claim.

Next you’re asked to read actual news articles that contain claims in PolitiFact’s database. In “highlight mode,” the claim in question is unobtrusively highlighted; click the text for PolitiFact’s evaluation. In “goggles mode,” all of the text that follows the first highlighted claim is blurred out, making it impossible to read on without engaging the claim first. In “safe mode,” all highlighted phrases are blocked out, forcing the user to reveal each one by one.

Finally, you’re shown many of the same claims again — this time without the goggles — and asked to rate their trueness. Schultz wants to see whether his software influences the way people process information and make conclusions.

“All of the current interface prototypes are walking different lines between theoretical effectiveness and convenience. I’m just as curious as you to see if the interfaces end up creating different degrees of impact in the study environment,” Schultz told me in an email.

I found that the latter two settings, goggles mode and safe mode, focused my attention on phrases I might otherwise have brushed past. Sure, my internal skepticism detector usually starts beeping whenever I see quotation marks, but this exercise also forced me to consider claims in a reporter’s copy — e.g., “About 2.5 million young adults from age 19 to 25 attained health coverage as a result of the Affordable Care Act…” — the kind of information I’m more likely to assume is true.

After using the goggles for awhile, it was impossible to read articles without a skepticism bordering on incredulity.

If you don’t want to wade through the study, there’s also a static demo page available. Schultz said the study and the demo are both “canned,” in that he preselected articles he knew to contain claims checked by PolitiFact. “Other than that, it’s using a live working version of Truth Goggles,” he said.

Schultz doesn’t want to talk much yet about the study design, lest he taint the data. “I’ll just say that the results so far appear incredibly interesting,” he said. His thesis, and a working version of the software, should be finished by as early as next week.

POSTED     May 14, 2012, 10:03 a.m.
SEE MORE ON Audience & Social
Show tags
 
Join the 60,000 who get the freshest future-of-journalism news in our daily email.
Deepfake detection improves when using algorithms that are more aware of demographic diversity
“Our research addresses deepfake detection algorithms’ fairness, rather than just attempting to balance the data. It offers a new approach to algorithm design that considers demographic fairness as a core aspect.”
What it takes to run a metro newspaper in the digital era, according to four top editors
“People will pay you to make their lives easier, even when it comes to telling them which burrito to eat.”
Newsweek is making generative AI a fixture in its newsroom
The legacy publication is leaning on AI for video production, a new breaking news team, and first drafts of some stories.