Dhe field of artificial intelligence (AI), which is not lacking in hype, has one candidate richer: ChatGPT, a speech synthesis system that generates texts of remarkable quality. In the debate to date, the limits of current technology have rightly been named, for example by pointing out the superficiality of many texts generated by this AI or by demonstrating that the system can be prompted to make racist, misogynistic or otherwise inadequate statements despite existing specifications. On the other hand, it is also rightly recognized that there is enormous potential for productivity increases here, especially if expectations remain realistic and the generated texts are seen as the basis for human revision and not naively as the end product of a creative process. From this perspective, ChatGPT and similar language AI models are further assistance systems for humans, perhaps comparable to machine translation, a spell checker or even the manual Internet search that inspires us every day.
The amazing quality of the texts generated by ChatGPT leads to many questions. The handling of basic examinations, for example in schools and universities, which could be generated via ChatGPT, is currently enjoying the greatest attention. The writing of journalistic texts is also re-examined. It can be assumed that the performance of the technology will continue to increase, so that it will soon be able to be used in many other, very different and complex fields of application. That’s why we should think now about how we want to deal with the technology and its possibilities in the future.