After the creation of ChatGPT, many students were tempted to appropriate AI-generated texts and present them as own original contribution. Therefore, professors all around the world are skeptical of integrating large language models into their courses because they fear that they will additionally increase academic dishonesty. After the professors of the Computer Ethics course, whose goal is, among other things, to raise the ethical standards of students and increase their academic integrity, noted massive cheating in academic writing at the end of 2022, they prepared a strategy for the realization and delivery of assignments that explicitly shows where and how used large language models. Students applied this approach for producing two group essay assignments during the winter semester of this academic year. This paper explains the approach in detail and, based on the experience with a group of over 150 students, evaluates its impact on essay writing, stimulating the responsible use of technology and improving the quality of delivered assignments. Based on extensive observations of the use of artificial intelligence in writing and personal impressions of students, this paper offers recommendations on how various applications of large language models can be used to improve student outcomes without encouraging academic dishonesty.