Work

Should Artificial Intelligence Align With Human Thinking In Innovation Processes?

Year: 2026
Type: article
Abstract: 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 assumptions and provides alternative framings to promote divergent exploration and an implementation chatbot that probe feasibility and accelerates convergent execution.We pilot the system in two innovation focused hackathons to show that human-thinking-aligned assistance preserves idea diversity during ideation while facilitating efficient convergence toward feasible prototypes during implementation.We plan to conduct a large-scale field experiment.
Institution: ETH Zurich
Language: English
Open Access status: green
OpenAlex ID: W7163852094
FWCI:
Cites: 9
Cited by:
Related to: 10