project-design reflection-questions
“Conducting research, especially with complex and often opaque ML models, confers a degree of power in the researcher. The researcher decides which technical features will be used, and which data flows will be enacted. This paternalistic approach raises questions about the extent to which the stakeholders or data subjects are aware and in agreement with the experiments or data collections conducted in their social domains.26 Additional to questions from IRE
1. 0, other questions are therefore relevant to highlight: - How does the researcher justify the foreseen intervention in a social, economic, or legal context by technical means? - Is the researcher able to understand and explain in an accessible manner how AI is to be used in the research?” - Is the researcher able to explain why the approach by technical means and the use of AI are better suited than any alternative methodology? - Have the persons who will be affected by the AI system requested the research? - If not, have they been informed and have they agreed? - If gaining the informed consent of all data subjects is infeasible, can the researcher obtain proxy consent from a representative or institutional ethics board on their behalf? - How has the balance between advantageous ends and individual freedom been struck? - Which values did the organization decide to promote, and how?"(franzke et al., 2020, p. 37-38)