Equitable Computing and AI

The Equitable Computing and AI group’s aims are: to raise awareness of how AI could impact equality; to proactively facilitate conversation in this space; to research the challenges that currently exist or may develop; and to be thought-leaders on solutions and the standards that practitioners should uphold.

FCA's EQUITABLE COMPUTING AND AI GROUP

As the pervasiveness of computing expands and AI seeps into the structures of society, it creates a hybrid of human-machine decision-making that affects us all.  Bias threatens to impact everything from access to capital to our criminal justice system.  Automation and machine enabled decision-making will impact the fabric of society. Mounting inequality is a concern. Transparency and accountability is vital. We must address unequal exposure to risk and to opportunity. We must ensure equitable computing that is humane, fair, and inclusive – for all.

The goals of this working group are to raise awareness of how AI could impact equality, to proactively facilitate conversation, to research the challenges that currently exist or may develop, and to be thought-leaders on solutions and the standards that practitioners should uphold.  The overarching aim is to maximize the benefit and opportunity offered by computing to all and to ensure fairness, representation, explainability, and reproducibility of methods.

More specifically, The Equitable Computing & AI Working group has three primary goals:

Educate the broader public about Equitable Computing & AI

  • Distribute accurate, understandable information to the public.
  • Raise awareness of equitable computing issues by partnering with visible/vocal folks in AI (with an additional special focus in developing countries).
  • Help people understand how their data is used.

Spread best practices and generate tangible technical artifacts for direct use by practitioners

  • Understand the baseline – what are today’s best practices? Potentially publish a paper and/or a series of blogs.
  • Work cross-functionally to create an Apache-style AI and Equitable computing “License” that can be attached to source code/product licenses.
  • Work with existing bodies working in the space including regulatory bodies, ISO, research working groups.

Elevate and promote interest in AI for Good

  • Identify ways of promoting projects, research and companies that are proactively working on AI for good.

Group Members


CHAIR:
Julia Haines
juliahaines@google.com

CO-CHAIR:
Danish Contractor
dcontrac@in.ibm.com