A frequently neglected area of software engineering is documentation. In this article, we take a closer look at how the trend towards faster release cycles can be reconciled with software documentation by applying documentation as code and artificial intelligence. The article focuses on the automation of documentation close to code and shows where generative AI can be useful here. The potential and challenges arising from the combination of documentation as code and artificial intelligence are discussed and demonstrated with practical examples using GitHub Copilot. In addition to documentation in parallel to development, legacy code documentation is also considered in this article. It brings additional challenges like bad quality or missing design documents. The correctness and completeness of the documentation will also be addressed and possible improvements to current approaches will be discussed.