Coursera Flash Sale
40% Off Coursera Plus for 3 Months!
Grab it
Explore how generative AI tools can better support creative decision-making and iterative design processes in this computer science colloquium lecture. Learn about the fundamental challenges creators face when moving between different levels of abstraction in content creation, from high-level conceptual decisions to precise technical details. Discover why current AI tools function as unpredictable black boxes that force users into trial-and-error workflows rather than supporting meaningful design exploration. Examine proposed features that generative AI systems should incorporate to enable true creative collaboration, including output consistency across iterations, hierarchical task decomposition, and support for rapid, reversible actions. Gain insights into potential approaches for building more predictable and controllable AI tools, along with demonstrations of prototype implementations developed at Stanford that embody these principles. The presentation addresses the gap between AI's content generation capabilities and the nuanced, iterative nature of human creative processes, offering a vision for more effective human-AI collaboration in creative domains.
