Artificial intelligence for testers: is it a threat or an assistant?

On June 15, an informative and inspiring webinar with experts from EPAM Ukraine was held on the role of artificial intelligence in the work of testers. The growing interest in artificial intelligence (AI) in various fields raises many practical and ethical questions, the answers to which are yet to be found. One of the questions we face is: does AI threaten the existence of the position of a tester and what to do about it?

During the presentation, the speakers discussed this topical and quite controversial issue, and shared their forecasts with the audience. The following experts presented during the webinar: Natalia Bohushevska (Software Testing Manager at EPAM) and Oksana Kasianenko (Lead Software Test Automation Engineer at EPAM).

In the beginning, Natalia reviewed the typical steps in the work of a manual tester and reminded about the following main testing phases: analysis of requirements, planning, development of cases, setting up test environments, data preparation, execution of tests and collection of results. Then, she spoke about the important AI principles and capabilities that should be used in the testing process. The speaker simulated a situation to show the importance of correct formulation of prompts and provision of context for ChatGPT. Natalia also shared useful ideas on how artificial intelligence can be used to optimize testing processes.

In the next part of the webinar, Oksana talked about an interesting tool, namely the GitHub Copilot plugin, and explained how it can be utilized in the process of building the Test Automation framework. The speakers discussed Copilot, the main scenarios of its use and its operation in practice.
Oksana shared statistical data from the official GitHub site, which indicates an increase in development productivity due to the use of artificial intelligence.

Oksana and Natalia also discussed ways to communicate with Copilot and utilize it for development, refactoring, and documentation purposes. In the end, the speakers shared information about Copilot limitations and answered the questions of the webinar participants, such as:

• examples of using Copilot for real projects
• whether Copilot can steal the code and how to deal with it
• whether Copilot can offer complex solutions
• safety of ChatGPT use for the development of documentation
• the factors on which the duration of validation depends on
• whether artificial intelligence can be tested for errors.

It will be exciting for testers and those interested in testing to watch the webinar and learn about the prospects for the future development of artificial intelligence in this area. The full webinar recording in Ukrainian can be found at EPAM YouTube channel.