AI Ethics Guide

The field of AI is constantly changing. Because of this rapid change, ethics of AI today is more about ensuring the right questions are asked, rather than defining the right answers.

These principles and questions help us ensure our values are applied to the Artificial Intelligence solutions that we build.

• Transparency: How have we ensured that our AI systems are transparent and understandable to humans?
• Fairness: How have we made sure our AI systems are fair and unbiased and don't discriminate against any particular group?
• Privacy: How have we protected the privacy of individuals and their data?
• Accountability and Responsibility: How have we ensured clear accountability and responsibility for our AI systems and their output?
• Safety and Security: How have we guaranteed the safety and security of our AI systems and prevented them from being used to cause harm?
• Human Control: How have we ensured that our AI systems are always under human control and don't make decisions that cannot be overridden by humans?

Many of these questions can actually be asked of all of the products we build, the algorithms we use, and not just those that rely on AI.

We have an #ethics Slack room. We use this room to help each other answer these questions about the products we build and the tools we use.

We also recommend that our team members complete some AI Ethics training. The training enhances our understanding of this critical issue.