Algorithmic bias: Can we stop bad data before it starts?

Date & Time:
Wed, May 8, 3:30 pm to 4:10 pm
Track:
Insights and Inspiration
ID:
DIBS1709
Speaker(s):
Chris Bedi, CIO, Global, ServiceNow
John Medicke, Distinguished Engineer, Watson Talent, IBM
Sheri Feinzig, Director, IBM Talent Management Solutions, IBM
Description:

In the 1970s and 1980s, orchestras had a problem. Even the most acclaimed had 10% or fewer women. Over the past 30 years, however, they’ve made major improvements in gender composition by implementing one simple change: blind auditions. Putting a screen between orchestral judges and prospects caused musicians to be selected solely based on the quality of their performance.

This type of offline bias shares similarities with the growing issue of algorithmic bias in the digital world. Although artificial intelligence and machine learning increasingly influence how digital products and systems work and the results they recommend, they’re only as good as the data that goes into them. Bad data leads to bias.

Join this session to hear how industry experts and technologists are addressing bias in algorithms and the actions you can take to advance technology that is blind to bias.

Audience:
Session Type:
Diversity, Inclusion, and Belonging (DIBS)
Product(s):
Industry:
Topics:
Culture and Workforce
Apps/Capabilities:
Tags: