For years, the focus of ML work has been to push frontiers, develop new capabilities, and achieve increased accuracy on a variety of tasks. So, where do we go from here? Bigger, faster, more accurate? Or do we take a step back and dig into transparent, fair, and explainable?
Ethical AI is rapidly being forced on industry as a key concern. Yet despite the significant investments, organizations struggle to operationalize ethical AI. We realized that we had to start doing things differently. In this talk I will share Pandata’s journey to incorporating Ethical AI into our practice.
• Understanding what it means to be an ethical data science practice, and why it is easier said than done
• Learning to have hard conversations among ourselves and our clients about bias & risk, and to articulate the importance of these conversations as practical and not just academic
• Moving from ethics as a value to ethics as a virtue. How we went from ethics as an aspirational thing we say to an actual way of working
• Some tools and processes we've adopted along the way to help, including our hiring processes
The result is more fulfilled and representative data science teams and better than industry average retention while fostering pro-active risk management.
By attending this talk you will learn:
• Make the business case for building an ethical data science practice
• Move beyond platitudes to get to the real work of ethical data science
• Incorporate ethics into your talent acquisition and development strategies
Abstract & Bio
Building an Ethical Data Science Practice