With recent events putting a spotlight on anti-racism, social-justice, climate change, and mental health there's a call for increased ethics and transparency in business. Companies are, rightfully, feeling responsible for providing underrepresented employees with the same treatment and opportunities as their majority counterparts. AI can, and will, be used to help companies understand their environment, develop strategies for improvement and monitor progress. And, as AI is used to make increasingly complex and life-changing decisions, it is critical to ensure that these decisions are fair, equitable and explainable. Unfortunately, it is becoming increasingly clear that, much like humans, AI can be biased. It is therefore imperative that as we develop AI solutions, we are fully aware of the dangers of bias, understand how bias can manifest and know how to take steps to address and minimize it.
In this session you will learn:
*Definitions of fairness, regulated domains and protected classes
*How bias can manifest in AI
*How bias in AI can be measured, tracked and reduced
*Best practices for ensuring that bias doesn't creep into AI/ML models over time
*How explainability can be used to perform real-time checks on predictions