Access and Feeds

Open Data: AI Defends Open Data Against Malicious Graffiti, but is not yet Foolproof

By Dick Weisinger

Open Data projects like Wikipedia provide useful repositories of information.  But open data repositories are often targets of malicious edit attacks. Data is changed in ways to alter or deface information.

Three years ago, Wikipedia deployed AI to try to combat data tampering.  Wikipedia said that “by combining open data and open source machine learning algorithms, our goal is to make quality control in Wikipedia more transparent, auditable, and easy to experiment with.”

This past summer, Mapbox,  a large provider of custom online maps for websites and applications, was attacked and place names on maps of New York and elsewhere were altered to show anti-Semitic names. Mapbox too uses AI to try to eliminate these kinds of attacks.

Eric Gundersen, Mapbox CEO, told ABC News that “everything that’s quarantined by the AI system is reviewed. So 70,000 things a day are quarantined. In this case — and this is so disgusting –- this person made 80 edits all focused on anti-Semitic language from New York to New Zealand. Every single one of those edits was captured, every single one was reviewed. One of those edits was overridden and pushed live.”

Gundersen said that “It shows that no matter how smart you build your AI, at the human review stage you can still have a point of failure.”

Digg This
Reddit This
Stumble Now!
Buzz This
Vote on DZone
Share on Facebook
Bookmark this on Delicious
Kick It on DotNetKicks.com
Shout it
Share on LinkedIn
Bookmark this on Technorati
Post on Twitter
Google Buzz (aka. Google Reader)

Leave a Reply

Your email address will not be published. Required fields are marked *

*

five × 3 =