Design thrives on ambiguity—the best ideas come from challenging assumptions and navigating uncertainty. LLMs, however, are designed to fill in gaps with existing knowledge. Left unchecked, it reinforces patterns rather than expanding them.
This research explores how AI can expand, not limit, human creativity. Inspired by Doug Engelbart's vision of augmentation over automation, it explores AI-powered tools that help designers uncover blind spots, challenge defaults, and manage complexity.
AI's predictive nature poses three risks:
•
Bias
AI inherits societal biases, reinforcing dominant narratives rather than challenging them.
•
Misplaced trust
Often lacking the proper context, AI will hallucinate in plausible-sounding ways.
•
Feedback Loops
LLMs compress human-generated data, reinforcing the same patterns.
AI for Reflection
Rather than treating AI as a shortcut to solutions, this research argues for using AI as a tool for introspection and challenging assumptions:
•
Assumption Mapping
AI reveals implicit biases in problem definitions.
•
Divergent Paths
Instead of a "best" solution, AI generates radically different perspectives.
•
Tree-Based Exploration
A structured interface maps how ideas evolve, highlighting where biases may be unconsciously limiting thinking.
The goal isn't to make design more efficient but more effective— more rigorous, insightful, and human. AI should extend thinking, not limit it.