Andrew Oram

Much hand-wringing appears in the press about the seemingly unstoppable ascendance of a few large corporations in computing. Everything seems to be increasingly centralized in such corporations (a trend popularly called the "cloud," although Richard Stallman has repeatedly criticized the use of that buzzword). This presentation will explain why such centralization and the triumph of first movers is facilitated by three technological factors: the end of Moore's Law, compiling complex algorithms into hardware (which may reach its climax in quantum computing), and the value of aggregating large amounts of data.

Algorithms are the new boogie men in social control and institutional discrimination. There is bias and lack of accountability in the algorithms that determine who gets hired for a job, who can get a loan, who qualifies for insurance, and even who goes to jail.

Well-designed algorithms can eliminate natural human bias. But with black-box algorithms, humans seem to be losing control over the machines that control our lives.

Sharing the source code implementing algorithms isn’t enough. Bias may be built into algorithms: for instance, an algorithm using actual random stops and arrests could recommend harsh treatment for blacks, as they are targeted more frequently by cops.

Research suggests a counter-intuitive approach to ameliorating bias. One must not be blind to demographic categories who experience discrimination--instead one must actively monitor these factors.

Panelists will present their views for a few minutes, then taking comments and questions from the audience.

Other people: Ifeoma Ajunwa Geoff A. Cohen Ben Green