Union Square Ventures: "The Ethics of Algorithms"

From USV:

Last week we hosted 30 Data Analysts, Engineers and Product Managers to share stories, thoughts and ideas on data challenges they’re currently facing. One theme that came up in multiple conversations was around the ethics of machine learning and recommendation engines.

Recommendation engines and algorithms create positive feedback loops that reinforce a desired behavior by replicating the conditions that led to that outcome. These tools can provide a great user experience by recommending new content to you before you even went to search for it. Maybe Amazon recommends you a new book by an author you’ve already purchased books from. Facebook may suggest a friend to connect with based on your connection to 10+ mutual friends. The recommendations are there to help make the product more useful, personalized and easier to find value.

Issues arise with these algorithms when they clash with ethics and morals. What if a Meetup group only suggests technical talks to men because the majority of members are also men, so they are more likely to join that group? What if an algorithm set up to fight fraud actually buckets some people incorrectly, making it impossible for them to transact on certain websites?...MORE

HT: USV's Fred Wilson on his personal blog.