On 30 November 2022 OpenAI released ChatGPT to the public and in only two months it had reached 100 million users. ChatGPT is far from the first generative AI tool. Others close to our topic are for instance GitHub Copilot, Google Bard, ChatSonic and YouChat. However, due to its simple graphical user interface and its ability to answer question in relation to programming tasks, understanding JavaScript amongst others, ChatGPT struck us like lightning. Students, teachers and researchers immediately started to use and explore its capacities. In other words, students adopted it. However, according to our experiences the adoption was often without thinking on the consequence of using it, not only in relation to the perspective of learning programming but also in relation to when it's okay to use the tool and when it’s not. During the spring and the following autumn, the authors therefore started to raise and discuss mainly four questions. 

1. Do the students know how to interact with ChatGPT? 

2. Do the students formulate and ask the “right” questions to ChatGPT? 

3. Do the students reflect upon the dialog with ChatGPT generally or what they are typing into the user interface in ChatGPT specifically? 

4. Do the students reflect upon what the ChatGPT tool actually returns (answers) on a specific question and/or dialog? 

In this presentation we will present our experiences in relation to two programming courses in spring 2023 in higher education one focusing on JavaScript for the client (the web browser) and one focusing on JavaScript for the server. We will both present challenges in the relation to using AI generally and ChatGPT specifically and how we solved some of the challenges including what did work and what did not work in both courses. Finally, we will present some thoughts about future usage of AI services, such as GhatGPT, in relation to not only programming coursers generally but also design and evaluation of examinations and detection of misbehaviour including cheating specifically.