Should Artificial Intelligence Align With Human Thinking In Innovation Processes?

Large Language Models (LLMs) are often integrated into today's innovation processes, but uniform AI assistance is incongruent with innovators' shifting cognitive states, encouraging over-reliance and premature convergence that erode idea diversity.We introduce a phase-specific, multi-agent human-AI innovation system that offers distinct AI assistance across exploration (divergence) and implementation (convergence).The system augments innovation phases with ideation support that probes assumption