4. The Ethics of AI Data Storage

Overview

In this chapter, you will learn about how ethics relate to Artificial Intelligence (AI) by looking at some of the largest industry scandals where AI ran up against morality. We will pe deep into several case studies, examining everything from AI being used to manipulate elections to AI displaying racial and sexist prejudices. We'll implement a simple sentiment classifier to differentiate between positive and negative words and sentences. We'll observe how this works in many cases and display the problematic biases and human stereotypes in the classifier. We'll gain hands-on experience with word embeddings and see how word embeddings can be used to represent how certain words relate to each other.

By the end of this chapter, you'll have learned how to evaluate the ethical aspects of the AI systems you build. You will know how to examine the potential consequences of using AI-based systems, and you'll know where to pe deeper to look for potential ethical problems. You'll be able to identify appropriate problems where word embeddings might be useful, and you'll be able to use Python's spaCy library for some common natural language processing (NLP) tasks.