course
10/10/2023 – 10/02/2024
Film University Babelsberg Konrad Wolf
Technological developments in computation, data, algorithms, and especially in recent times, systems of machine learning have led to an increasing number of tasks, which were until recently exclusively performed by humans, now being undertaken by machines. In the realm of social media, fake news and troll factories have already demonstrated how profound the effects of these systems can be on democracies, societies, and our understanding of truth. With regard to machine learning, the first effects on schools, universities, the workplace, and social interactions are already emerging. A particularly relevant topic in this context is explainable artificial intelligence. Many machine learning methods resemble a black-box process; how exactly the systems make decisions is not directly discernible from the system itself. Not only are the outcomes of systems like ChatGPT, Stable Diffusion, Midjourney, or Dall-E uncertain, but also their inherent biases and worldviews, which however are reflected in every query.
At this juncture, the course aims to employ artistic interventions to better fathom these systems. The fundamental idea is to use "prompting" – the text input into the system – to gain a sense of the inclinations, basic attitudes, and errors of the systems, and to analyze, reflect upon, and communicate them. The results can range from texts (How to Read an AI Image), short films (Hardly Working) to interactive web applications (Artificial Worldviews).