Language models have been around for a long time. But the release of ChatGPT hit like a bomb. Maybe that’s because it’s so easy to access. I think everyone now realizes that AI is changing our world. Some people even see greater upheavals than the “invention” of the internet. People are experimenting with it today, there are no limits to creativity and you can learn from others, because experiences are readily shared - worldwide. But how can we, as testers, make ChatGPT and co. work for us?
“Isn’t that fascinating? I feel like I did 40 years ago as a teenager in those science fiction novels: you have an artificial intelligence that you can talk to!” - Klaudia Dussa-Zieger, Michael Heller
Klaudia Dussa-Zieger has been in charge of software testing and quality assurance at imbus AG for over 25 years and has been the first female president of the ISTQB since 2023, as well as chairing the DIN working committee for systems and software engineering since March 2009. Her focus is on test management, process improvement and international standardization.
Michael Heller is a consultant at imbus AG with over 15 years of experience in the introduction of standard software and agile methods, specializing in training, scaled agility and test process optimization. He is an ISTQB trainer, certified Scaled Agile Framework Practice Consultant (SPC) and co-author of the ISTQB curriculum ATLaS - Agile Test Leadership at Scale.
Highlights of this episode:
Today I’m talking to Klaudia Dussa-Zieger and Michael Helöler about the many possible applications of ChatGPT in the field of software testing. From generating test cases to creating scripts and increasing creativity in the testing process, the two show how ChatGPT can improve efficiency and open up new perspectives.
The world of software testing is undergoing significant change, driven by advances in the field of artificial intelligence (AI). In particular, ChatGPT, an advanced language model, has piqued the interest of testing experts. In our latest podcast episode, we take you on an exciting journey and explore the many ways ChatGPT can enrich the process of software testing. Klaudia Dussa-Zieger and Michael Heller share their experiences and insights on the use of this technology in real projects.
Klaudia and Michael begin their exploration of ChatGPT with a playful curiosity. They emphasize the importance of approaching the topic without hesitation and simply getting started. Using their own experiments, they illustrate how even testers without in-depth knowledge of neural networks or data science can benefit from this technology. They emphasize that it is crucial to ask the right questions and critically evaluate the answers.
The versatility of ChatGPT is underpinned by various use cases. From generating test cases from requirements specifications to supporting the creation of test specifications - the possibilities seem endless. Klaudia particularly emphasizes how impressive it is that ChatGPT is able to deliver comprehensive and creative test ideas. But she also warns of the limitations: You have to know when ChatGPT is telling ‘total garbage’.
Although the potential uses of ChatGPT are impressive, there are also challenges and limitations. One key challenge is the governance issue: where does the data reside and who owns the results? Klaudia shares her concerns about the use of project-specific data and the need for clear regulation in companies.
Looking to the future, Klaudia and Michael are optimistic that the use of AI technologies such as ChatGPT will continue to grow in importance. They hope to increase awareness of other AI solutions in the testing sector and emphasize that it is important to stay on the ball and not shy away from experimenting with new tools.
ChatGPT has the potential to fundamentally change the field of software testing. Thanks to its versatility, it can not only increase efficiency, but also open up fresh perspectives. Klaudia and Michael impressively illustrate how important it is to be open to new technologies and to actively experiment with them. Their experiences clearly show that we may be at the beginning of a paradigm shift in software testing.