Stay on Top of Enterprise Technology Trends
Get updates impacting your industry from our GigaOm Research Community
Are computer algorithms making you dumber? Yes, says Eric Berlow, founder of Vibrant Data Labs. Speaking at the GigaOM Structure: Data 2013 conference on Wednesday, Berlow offered several compelling examples of this phenomenon as well as an approach to augment algorithms with more human input.
“There’s lots of content in the newspaper,” Berlow noted. “After viewing the most-emailed stories for a few weeks, I asked myself, where did all the news go?” Think back to the Presidential debates, Berlow said. If you focused solely on topics provided by news algorithms, you’d be reading nothing but stories about Big Bird and binders full of women.
Amplifying crowd behavior is a start when it comes to managing societal data but we need to flip the approach, Berlow said. “How do we harness and amplify our collective creativity to solve the world’s problems?”
To do this, we need to find crowd-sourcing solutions that aren’t just the sum of parts, but are greater than the sum of parts. Photocity is a good example, according to Berlow. It takes user submitted 2D camera images and creates 3D images from them; a product that didn’t exist until the crowd’s data was assimilated.
This leads to one of the biggest challenges of our time with data: The personal data problem, where you are both the customer and the product. How can we spark a new personal data economy?
Through the WeTheData.org project, Berlow offers a suggestion, finding how all of our personal data is interconnected. By gathering human input first on approximately 90 personal data challenges and mapping this complexity, the project determined that the top problems emerging are digital access, digital trust, data literacy, platform openness.
This may sound obvious in retrospect, Berlow said, but “so too was gravity after we discovered it.” And now that the personal data economy problem is better defined, algorithms can be applied to focus on the biggest issues, not every single one.
Check out the rest of our Structure:Data 2013 coverage here, and a video embed of the session follows below:
A transcription of the video follows on the next page