AI tools such as ChatGPT and Copilot recently have become available to everyone, including students learning how to program. Generating code to solve a specific problem, or asking for help when your code doesn’t work, has never been so easy and within reach. Discussions on whether we should prohibit or embrace these tools are prevalent. While everyone has some anecdotal evidence of the usefulness and/or problems regarding these tools, researchers all over the world have started to study this development more methodically.
In this talk I will give an overview of the research studies that have appeared about these tools in the context of computer science and learning programming. At the 'Software Technology for Learning and Teaching' research group at the department of Information and Computing Sciences at Utrecht University we have also started studying several aspects of these AI tools, for which I will show some preliminary results.