Considerations About Integration of GenAI into Products and Services from an Ethical and Legal Perspective

Generative Artificial Intelligence (GenAI) has gained significant attention since the end of 2022. In many existing products and services chat-bots are integrated and completely new services based on GenAI, particularly those based on Large Language Models (LLMs) are under development or have been recently released. However, many aspects regarding the use of LLMs remain unclear, particularly in relation to the training and the inference phases. Training-related questions are about the selection of the data and their legal constraints and ethical considerations. Trained models are offered under new specific "AI licenses" with implications on the context of usage. Anyway, there are open aspects about inference outcomes respective with respect to the usage of LLMs. This work examines considerations in the context of the product and service development for the generation of test-cases facilitated by a LLM. Test-case generation is used in Information Technology (IT) to make derivations from specifications for test cases (respectively test-scripts) to validate the implementation. Test-case design and implementation can potentially be facilitated by LLMs - based on a hypothetical case, relevant aspects and considerations are discussed also including legal aspects based on the perspective of German law.