T.A.R.G.E.T. – Testing with AI Reinforced GUI Embedded Technology

Software testing via graphical user interfaces, GUI testing, is a common but costly practice in industry that contributes to the global cost of trillions of dollars each year for software quality assurance. GUI testing is required to ensure the correct behavior of a system but has several challenges. The project will look at AI-based GUI testing.

Aim of the project

Graphical User Interface (GUI) software testing is a common but costly practice in industry. Graphical User Interface (GUI), or GUI testing, is a common but costly practice in industry that contributes to the global cost of trillions of dollars each year for software quality assurance. GUI testing is required to ensure the correct behavior of a system but has several challenges. These include technical knowledge requirements for developing and maintaining test scripts and lack of tool support for automating manual tests. These challenges prevent domain experts, such as end-users, from contributing to the testing process and therefore from having their needs met.

The T.A.R.G.E.T. project aims to address these challenges by creating a new and easy-to-use AI-based approach for GUI testing that combines Augmented Testing with Generative AI. Augmented testing is used to record and visualize test data for the tester on the software's graphical interface while AI agents analyze the recorded data, learn from it, and then provide the tester with suggestions for new test cases and test data to cover all user needs and perspectives. Testers and AI thus learn from each other in a concept we call human-machine symbiosis.

The project is carried out in a consortium of experts in AI and testing with the aim of producing new knowledge for end-to-end AI solutions, new technology, and a tool for realizing human-machine symbiosis, an approach that will strengthen Swedish industry's ability to innovate and test advanced digital systems.