I'm still making my way back from Beijing, and have plenty of notes to write up for WWW 2008. However, I thought I'd point ahead to my next trip: The Future of News - a workshop being held at Princeton's Centre for Information Technology Policy. I'm really looking forward to this event as it ties in some recent work I've been doing at Microsoft (including supporting MSR's Blews team), some broader thoughts about the relationship between politics, broadcast media and social media, and my current (belated) reading of Lessig's Code (v 2.0).
Some thoughts starters:
- Main stream news is increasingly driven by simple economic forces. I suspect that the increase in personality driven news is due to a person's ability to forge emotional connections with a homophillic (read 'loyal') audience.
- Social media has demonstrated its ability to act as a fifth estate.
- More and more algorithmic editors are appearing (e.g. Google news, Silo Breaker, European Media Monitor).
- Despite claims to the contrary, algorithmic editors (that is to say, programs which automate the selection and presentation of news stories) do encode certain biases (witness the lack of Dalai Lama quotes in Google news' quotes feature) intended or otherwise.
- The designers of algorithmic news systems have an opportunity to do social good (e.g. acting against homophilly)
- Seeing data (visualization) is a necessary first step in understanding what could be done with it.
- Principles of data representation are required to help prevent blind siding automation to certain qualities or opportunities in the data.
- Context provides a user with greater choice in media consumption, but adds to the cognitive effort involved.
This is very interesting. A few general comments about algorithmic editors and potential barriers to mainstream adoption:
1) I believe most older audiences are still tied to 20th-century media consumption habits, which emphasize a limited number of personalities, brands, and choices. These habits have carried over to the Web, and can still be seen in people's tendencies to visit a limited number of trusted news sources (often established print and broadcast media outlets). I believe Google News, Techmeme, Silobreaker and other algorithm-driven sites are a hard sell for these audiences. The trust and familiarity is not there, and mixed in with familiar names are lots of unfamiliar and even despised names.
But the type of functionality that they offer, whether it be a computer-generated list of top headlines, or a visualization, might be more palatable if they are introduced by trusted brands -- imagine a CNN trends "widget" that appears on your Facebook profile, or on an article page on CNN.com, that tracks top items or contextually changes depending on the content of the Facebook profile or individual article?
2) An unfortunate aspect of the algorithmic editors is that they can be manipulated. Witness the rise of spam, SEO, and unusual items and sources appearing with high ranks in Techmeme and Google News.
Posted by: Ian Lamont | April 30, 2008 at 04:45 PM
Ian,
These are great points. I'm guessing that there are blends of approaches that may be more appropriate to gain adoption. In addition, there may also be different delivery mechanisms that provide more comfort, trust, etc.
Posted by: Matthew Hurst | May 01, 2008 at 10:32 AM