Abstract
AI artifacts are technological ones with AI as crucial components. Those actually used in practice are denoted as AI artifacts-in-practice. Intelligent behaviors of AI appear to be results of recognition and assessment of meaning of input data. Judgements or decisions by AI convey intended/perceived information to its user. This research first studies connection between the interpretive approach and intended/perceived information, and defines purposeful domain-specific AI (PDsAI) as well as AI-related terms. Based on it, we maintain that PDsAIs have been essential components of ICT systems since 1950s. Three essential factors which make AI ethics critical concern are clarified. We also propose ideas of supply-ethics and use-ethics, and insist on usefulness of the interpretive approach and the intended/perceived information aspect.
