Seite - 123 - in The Future of Software Quality Assurance

TestingArtificial Intelligence GerardNuman Abstract In AI, the algorithm is not coded but produced by a combination of training data, labelling (concepts) and the neural network. This is the essence of machine learning. The algorithm is not directly insightful and cannot be bug-fixed directly: it is “blackboxdevelopment”. AI systems are used in contexts with diverse data and usage. Choice in training data and labels brings risks in bias and transparency with possible high impact on realpeople.TestingAIfocussesonthese risks.AnAI testerneedsmoral, socialand worldly intelligence and awareness to bring out the users, their expectations and translate these in test cases that can be run repetitively and automated. AI testing includessettingupmetrics that translate test results inameaningfulandquantifiable evaluationof thesystem inorder fordevelopers tooptimize the system. Keywords Software testing · Softwarequality · Artificial intelligence · Machine learning · Test automation 1 Introduction The future is AI. It has entered our everyday lives and is being used by major companies all around the world. Adaptability of AI seems endless. And yet many doubts and concerns exist. For example, in the case of self-driving cars: liability in case of accidents, wobbly object recognition and complex interaction with unpredictablehumantrafficparticipantsareblockingwidespreadacceptance. Some possible scary effects of AI have already manifested themselves. AI algorithms can create or enlarge bias. Like in the case of the ethnic cleansing in Myanmar, where tens of thousands of Rohingya were killed and one million fled. Already existing ethnic tension was supported by the Facebook algorithm, which G.Numan Polteq Test Services B.V.,Amersfoort, The Netherlands ©The Author(s) 2020 S.Goericke (ed.), The Future of Software QualityAssurance, https://doi.org/10.1007/978-3-030-29509-7_10 123