« Back to blog

TEDx Boston: Context is King, So Let’s Build Better Filters

I’m at TEDx Boston right now and having fun so far. The most recent speaker was Jodi Beggs, a behavioral economist and lecturer at Northeastern. She gave a great talk about how context matters and how the details in which an item is presented shapes how you think about it. How you taste wine depends on how much it costs, what your friends have told you about a musician impacts your taste for their music, etcetera.

This is all crucially important, and Beggs concluded by stating that she is bringing it all up in the hope that we might, with more awareness, be a bit more rational. And that’s a worthy goal. But we need to go further and build technology around these insights.

We need to build smarter filters that take these biases into account. Lots of online businesses — Yelp, Pandora, BostInno — function as filters in various realms. And those filters can be architected to help account for our biases, to shape context in a way that achieves the right goals.

A smart filter for music might seek to prioritize independent musicians, wine stories might include reviews done by blind taste tests, etcetera. Perhaps those are impractical; I’m not expert in either. But in the case of media (particularly political media), and the nasty problem of confirmation bias, we can experiment with smarter filters that help de-bias us.

I wrote about this for The Atlantic last year:

Context can affect bias, and on the Web — if I can riff on Lessig – code is context. So why not design media that accounts for the user’s biases and helps him or her overcome them? …

…as we design ever richer media experiences for the Web, we should pay attention to this kind of research and consider how we might create media that draws on it to counter our political biases. Our ability to reason is flawed in predictable ways. And as we increasingly link our social graph into our media experience, there is plenty of relevant data to mine for hints of bias. Why not improve our reasoning using cognitively sophisticated media?

You can read the whole piece for specific suggestions. My point now is to follow-up on Beggs’ great talk by pushing it a bit further. Technologists need to understand behavioral economics and the science of bias, and use that knowledge to build better filters.

from BostInno http://bostinno.com/2012/06/22/tedx-boston-context-is-king-so-lets-build-bett...