Tuesday, November 22, 2011

Then I don't know Arkansas

Can it be? The return of Spinoculars, only now with added Science®?

Bull beware: Truth goggles sniff
out suspicious sentences in news

You’re reading a wrap-up of the Sept. 22 Republican presidential debate when you land on this claim from Rep. Michele Bachmann: “President Obama has the lowest public approval ratings of any president in modern times.”

Really? You start googling for evidence. Maybe you scour the blogs or the fact-checking sites. It takes work, all that critical thinking.

That’s why Dan Schultz, a graduate student at the MIT Media Lab (and newly named Knight-Mozilla fellow for 2012), is devoting his thesis to automatic bullshit detection. Schultz is building what he calls truth goggles — not actual magical eyewear, alas, but software that flags suspicious claims in news articles and helps readers determine their truthiness. It’s possible because of a novel arrangement: Schultz struck a deal with fact-checker PolitiFact for access to its private APIs.

You can call it "truthiness" all you want, but it's still going to rely on someone else doing the work first -- if some semblance of the phrase is in the magic database to start with, and it's close enough for your algorithm to recognize it, and if you really need a bullshit detector to tell when Michele Bachmann is lying.

I don't think we ought to blame the grad student here, but the people who set out to write about what grad students do might want to consider recalibrating their credulous featurization machinery.


Blogger slifty said...

But recalibrating isn't nearly as catchy ;)

To be clear this isn't a spin detector: it's a critical ability activator. Sure, I will rely on humans for many steps of the way (and what's wrong with that?) but the real attempted contribution isn't in the algorithms, it's in the interaction.

Designing an interface to make you think about what you are reading (rather than creating something that you ignore / something that still allows you to fall back on natural tendencies towards confirmation bias) is actually the biggest challenge here.

11:08 PM, November 22, 2011  
Blogger fev said...

I appreciate the comment and, I hope, the point. But I'm reminded of my reaction to an Apple ad I just saw on the teevees: If you need an iPad to make you stay up all night reading about dinosaurs, ur doing it wrong.

12:05 AM, November 23, 2011  
Anonymous Anonymous said...

You actually don't "need someone else to do the work". This sort of thing is very close to the sort of question-answering work that the AI community has been doing for a couple of decades now (and has actually gotten pretty good at).

12:25 AM, November 23, 2011  
Blogger slifty said...

Point appreciated as well Fev -- Unfortunately, though, given the amount of misinformation that is spread and believed, I think it's safe to say that many people ar doing it wrong ;)

12:48 PM, November 23, 2011  

Post a Comment

Links to this post:

Create a Link

<< Home