This study analyzes and interprets image of humans in normative artificial intelligence (AI) ethics texts. The research material comprises six texts published from 2017 to 2021 for steering the development and use of AI systems toward human values. The study methodology, i.e., socio-ethical content analysis, is based on the conceptual distinctions and connections of action and good, and on the notion that the goals of different normative action types (promoting, maintaining, protecting, preventing, and prohibiting) imply their value (good or non-good). First, explicit mentions of the human nature and human beings are examined. Second, implicit images of humans are analyzed by reading the most frequently mentioned human goods (well-being, education, autonomy, responsibility, trust, and safety) in the context of normative action types. Third, goods and non-goods in AI ethics texts from earlier studies are interpreted to form six elements (epistemic, active, moral, relational, flourishing, and enabling) for a good human life, which depict an ideal person in the AI era. Finally, images of humans and ideal humans and their implications are discussed based on earlier research.