Between Adoption and Doubt: A Census of Moral Attitudes towards Artificial Intelligence among Students at the University of Lucerne

Executive summary:

1) As early adopters and key carriers of social activism and innovation, university students are a particularly important group for understanding the social ramifications of AI-based technologies. To get a more accurate picture of their use, perceptions, and moral evaluations of AI, we conducted a full census of students at the University of Lucerne(Switzerland). From May to June of 2023, 440 students from allfaculties participated in a standardized online survey.

2) Students are highly interested users of AI-based technologies. 98 percent of respondents have experiences with intelligent translators and 70 percent with chatbots, including the popular example of ChatGPT. Respond-ents are primarily interested in the moral consequences of new technologies, followed by legal frameworks, personal use, and economic consequences.

3) Despite the widespread use and interest, we observe high levels of uncertainty. The majority of respondents is undecided whether to trust AI or not. Additionally, only about every fifth respondent feels confident about their knowledge of AI. Thus, when it comes to AI-based technologies, students do not correspond to the widely employed figure of a digital native.

4) In general, students are highly skeptical of the future impact of artificial intelligence on society. Around 60 percent believe AI is going to take away jobs and reduce privacy. In contrast, only 22 percent believe AI will improve society. The largest disagreement concerns the question whether advanced AI poses a threat to the self-esteem of human beings.

5) A substantial number of students would still be willing to include AI in solving moral and political problems. One in four students thinks that AI is more likely than the Swiss population to be able to recognize optimal po-litical decisions. Every fifth student would ask AI for moral advice. At the same time, students are reluctant to attribute moral responsibility, moral worth, or moral rights to AI. This raises new questions forthe distribution of moral agency and accountability among human and artificial agents.

6) We derive various practical implications from these findings. Universities should enable students to assess the trustworthiness of AI and foster skills to understand and explain these new technologies. They should pro-vide spaces to discuss the morality of AI, especially employment effects, environmental impacts, and human self-esteem, while avoiding extreme predictions and alarmist accounts of the future of AI, whether dystopian or utopian.