Recently, safety-critical systems driven by AI technologies are expected to penetrate into society, there is a really call for methodologies of safety arguments or agreements. In this presentation, we introduce GSN/D-Case with its background and basic ideas and show some case studies.