Civicist

CIVIC TECH NEWS & ANALYSIS
Categories
#PDF Algorithms Civic Tech

CATHY O’NEIL ON WEAPONS OF MATH DESTRUCTION

CATHY O’NEIL ON WEAPONS OF MATH DESTRUCTION

Weapons of math destruction are characterized by their opacity, their power, their widespread use, their poor definitions of success, and their engendering of pernicious feedback loops.

  • Personal Democracy Forum is next week, and we’re reaching out to some of the speakers for a quick preview of their respective talks and panels. What follows are a few words from Cathy O’Neil, who writes at the blog mathbabe.org and is working on a book about the dark side of big data. O’Neil will deliver a talk on “Weapons of Math Destruction.”

    You’ll be speaking at the conference on the subject of weapons of math destruction. Give us a preview: what the heck are weapons of math destruction?!

    They are mathematical algorithms that are being deployed to make important life decisions for certain people at certain moments. They are characterized by their opacity, their power, their widespread use, their poor definitions of success, and their engendering of pernicious feedback loops. I will give a bunch of examples of WMD’s from education (the Value-Added Model for Teachers), the criminal justice system (evidence-based sentencing models), and politics (micro-targeting).

    The theme of the conference this year is the future of civic tech. As briefly as you like: Where do you think civic tech is going, what do we have to look forward to, and what pitfalls should people working in this sector be aware of?

    I’d say that my example with micro-targeting in politics is more or less an intersection of WMD’s with civic tech. I am, in other words, a civic tech skeptic.

    I’m focusing on the pitfalls. Civic tech has a lot of positive vibes but successful data work, which is usually done in the quest of power, money, or both, should teach us a few lessons. If we want data or technology to work for the public good, we have to make it so in a deliberate and thoughtful fashion. It’s not good enough for us to “open up the data” and wait for the tide that lifts all boats.