AWS Certified AI Practitioner (AIF-C01) Study Guide

Chapter 8. A Framework for Responsible AI

In 2014, Microsoft launched Xiaoice, an AI-powered chatbot in China that successfully attracted over 40 million users. A year later, Microsoft released a version on Twitter called Tay. The launch was a disaster. Tay quickly began spouting racist and hostile comments, forcing Microsoft to shut it down within 24 hours.

The incident occurred because Microsoft had a false sense of security from its experience in China, where stricter content limitations were in place, and underestimated the freewheeling nature of Twitter, where users actively tried to manipulate the bot. In a blog post, a Microsoft executive acknowledged the company had learned valuable lessons, stating that the challenges of AI are “just as much social as they are technical” and that caution is required when iterating in public forums.

This event spurred Microsoft to create its own principles for responsible AI—a framework for developing and deploying AI systems in a safe, trustworthy, and ethical way. As AI becomes more integrated into critical areas like healthcare and criminal justice, this focus on responsibility has never been more important and is a key part of the AIF-C01 exam.

Risks of Generative AI

As powerful as generative AI can be, its capabilities also introduce a wide range of risks—some subtle, others deeply disruptive. From producing toxic content to infringing on intellectual property and contributing to job displacement, these risks can have serious societal, ...