Artificial Intelligence and Software Testing: Building systems you can trust

by Rex Black, James Davenport, Joanna Olszewska

Independent Press Awards 2023 - Technology AI presents a new paradigm in the development of software, representing the biggest change to how we think about quality and testing in decades. Many of the well known issues around AI, such as bias, manifest themselves as quality management problems. This book explores AI from that angle and is aimed at testing and quality management practitioners who want to understand more, starting with the relationship between AI and trustworthiness, the issue of bias, and the complexities of testing machine learning systems. The book then pivots to cover how AI can be used itself in software test automation, before exploring the more advanced topic of ontologies, and how they can be brought together with AI and testing. Finally, examples of shift-right testing and AIOps are considered.

146 pages, Paperback

Published March 10, 2022

Reviews

[Catalin Negru · Rating 4 out of 5]

For testers who plan to learn for the AI ISTQB exam, this book is a good companion. However, unlike other books written by Rex Black & friends, which are dedicated to the exams and they complete the ISTQB courses, this book does not follow the structure of the AI ISTQB course. Instead, the book consists of 7 chapters, written by different experts, which focus on different aspects related to AI testing. These chapters cover only some of the aspects contained by the course. The reason for this is stated explicitly by one of the authors: AI testing is at the beginning of the road and many novelties are expected to show up in this regard in the near future.

[Piotr Lewkowicz · Rating 3 out of 5]

After reading "Artificial Intelligence and Software Testing: Building Systems You Can Trust," I’m not sure who I would recommend this book to. From the title, I expected a deeper focus on AI-related problems in software testing. The book starts well by discussing the challenges of checking AI's response quality, but I lost interest as I moved through the chapters.

Sometimes the content felt too complex for me, and other times too basic (like the testing pyramid concept). Because of this, I didn’t think it was the in-depth guide to AI testing I was hoping for. I might come back to this book after I gain more experience with AI, but by then, there might be newer books on the topic.

I did learn some new ideas, but I wouldn’t suggest this book to someone looking for a hands-on introduction to AI testing. I wanted more real-world examples of how to test AI systems.

The chapter by Adam Leon Smith was a highlight. However, some of the machine learning content was too advanced for me. This made me interested in checking out other works by Smith that might be easier to understand.

Overall, while the book has some good points, it didn’t meet my expectations. It might be more useful for readers who already have a solid grasp of AI and machine learning in software testing.