Trust Me, I'm a Data Scientist - Ethics for Builders of Data-Based Applications

Explore the ethical challenges in data science and machine learning through a practical example of building an AI-based virtual assistant for high school students. Delve into topics such as algorithmic fairness, model interpretability, and handling minority classes. Learn how unintended biases can infiltrate every step of the development process, even with the best intentions. Gain insights on identifying and avoiding major ethical pitfalls in the machine learning community. Suitable for beginner to intermediate data scientists and those working with data scientists, this 25-minute talk from the EuroPython 2018 conference requires no specific technical knowledge.

Introduction
Ethics
Why
Data collection
Bias
Encoding
Analogies
The problem
Building models
Interpretation
Minority classes
Validation
Cognitive Bias
Projection