In the case of societal security v. personal privacy

The transformative solution Fractal Responsible AI (RAI) framework

Foundational Principles

Encouraging Behaviors

Transparency All deserve white-box decision choice

Human-Centricity Evaluate biases

Every decision rule of an AI algorithm draws upon human-made assumptions. Moreover, the data used by the algorithm are not objective reflections but according to the context used to generate the data. A human dimension underpins our AI framework; hence the whole life cycle of AI created is governed by a sequence of principles and behaviors. We envision our AI as human-designed entities reflecting human values. By examining the present applications of AI, we have identified vital behaviors driven by principles that can help keep dangers posed by artificial intelligence at bay.

Contestability Newspaper test

Accountability Humans responsible

Explainability Decision factors traceable Adaptability Augment signals and update

Privacy & Safety Respect and protection from mis-use

Fairness Inclusion and equity as default

5

© 2022 Fractal Analytics Inc. All rights reserved

Made with FlippingBook - PDF hosting