Human vision emerges from the interplay of composition, pattern recognition, and emotional response. This process is deeply contextual, drawing on personal history, cultural frameworks, and embodied experience. Yet Machine vision operates in a different paradigm that is based on statistical analysis of vast datasets, reducing image complexity to mathematical relationships. While this approach may appear more objective, it carries its own forms of bias, as it encodes the perspectives and assumptions inherent in the data.All image-generating models are trained with a combination of systems that integrate both image and language processing, where visual and linguistic meaning intersect. We will explore the mechanisms of these vision models in generative AI through hands-on programming practices.In practice, we will work on small-scale models, including Ollama, CLIP, LLaVA, and others, experimenting with their capabilities to understand their potential, limitations, and biases. To be playful and experimental, yet critically engaged with AI as a tool in creative practice. We will read and discuss ideas from various perspectives and references, engaging with diverse voices, and contextualizing our practice within broader conversations of critical AI. Each technical exploration will be paired with critical reflection and artistic references on the implications and assumptions embedded in these systems.2. Terminal and small-scale Generative AI3. Working Mechanism of Image Generation4. Operational Image5. Text-Image Relations6. Finding the textual bias in the image system6. Multi-model Images as Cultural Snapshots7. Guest Workshop and Lecture by Artist Fabian Mosole