Improvements in AI Necessary to Counter Bias

The Issue: 

Experiments with AI-based recruitment find that automated recruitment discriminates against women and those with African-American sounding names.

The Cause:

White, male coders are inadvertently coding their bias into the algorithms.

Who is Trying to Solve This?

Proposed Solutions: 

  • CSR teams collect social impact data to counter potential AI bias
  • The Algorithmic Accountability Act which forces companies to audit AI for bias prior to use.
  • Companies can hire teams of computer scientists, sociologists, anthropologists, legal scholars, and activists to develop more fair and accurate data.

Harvard Business Review | Is AI Bias a Corporate Social Responsibility Issue?