CHAPTER 13Having Faith in the System: How Can We Trust the AI?
Prior generations of AI and ML technology focused on solving discrete tasks, whereas GenAI offers a wide range of diverse use cases and outputs. The downside of this versatility is a higher rate of error and risk. This chapter will explore how we measure and manage these machine-made outputs.
But first, Figure 13.1 shows a personal example from one of your authors (Andrea).
Andrea: It took me about seven or eight tries and about 20 minutes to craft the prompt to generate the image using OpenAI. It bothers me that the “1 - 2 - 3” is not on one line, but when I tried to make this small tweak it generated an entirely new image with other things I didn’t like as much. The stakes of this image not having the text bubble written out exactly as I’d like it here are obviously not that high. But when it comes to running your business and messaging your customers, precision and professionalism are a big deal. After all, the output represents you, and customers will judge you by it.
HOW EXTENSIVE ARE THE RISKS?
Jobs and Employment. The impact GenAI will have on the workforce concerns many observers. Some structural unemployment is inevitable—meaning, some jobs will be eliminated or changed—so the question becomes whether people can find new, meaningful jobs quickly enough.
FIGURE 13.1 Generating Creative Content with ...
Become an O’Reilly member and get unlimited access to this title plus top books and audiobooks from O’Reilly and nearly 200 top publishers, thousands of courses curated by job role, 150+ live events each month,
and much more.
Read now
Unlock full access