The performance of Large Language Models (LLMs) has evolved significantly in a short time. In relation to this, the pedagogical landscape has also changed, as more and more students adopt these technologies. In this article, we outline how the transforming landscape has affected academic cheating in higher education through the context of programming education. For this, we reflect on 3 different, easily available cheating methods utilizing large language models: generating answers to programming exercises, generating essay answers, and modifying existing resources, such as blog texts or Wikipedia articles, into new answers. For each method, we show simple examples and analyze the quality of the artifacts produced by the models and their usability as real assignment solutions. Moreover, we discuss the difficulties in detecting or preventing the illicit use of AI tools and the potential problems caused by such efforts. Finally, we try to predict how AI will shape the future of education.