Generative AI (such as ChatGPT, Bard, Midjourney, Dall-E, etc.) can create new content based on so-called prompts that people write to them in everyday language. These technologies have rapidly been developed and have proliferated recently, but at the cost of transparency and comprehensibility. It is generally challenging to discern where the texts and images originate from and why they appear as they do. HAIC-III aims to combine humanistic methods with art and design practices to explore Generative AI as an expressive, creative, and cultural interface. The project draws on interdisciplinary research fields, including digital literature, critical data studies, and software studies. HAIC-III analyzes how Generative AI is constructed as an interface and how it establishes new relationships with cultural archives. Additionally, HAIC-III experiments with designing alternative interfaces for Generative AI and employs artistic research to examine how AI impacts current societal and cultural perceptions. Fundamentally, the project investigates how we can use artistic and design-based approaches to address and manage problematic aspects of Generative AI use through a focus on creative human-AI collaboration. HAIC-III prioritizes human-AI collaboration over automation and views Generative AI as a cultural expression engine and creative interface.