Revolution or inflated expectations? Exploring the impact of generative AI on ideation in a practical sustainability context
通过与BSH家电集团合作，比较了AI（如ChatGPT）与传统专家研讨会在可持续包装创意生成中的效果，发现AI在新颖性上更优但可行性较低，并探讨了AI整合的障碍与建议。
The integration of generative AI (GenAI) into corporate innovation processes represents a significant shift in ideation methodologies. This study examines the comparative effectiveness of AI-generated ideation and traditional expert workshops. In collaboration with BSH Home Appliances Group (BSH), ideas were generated and evaluated using both expert-based and AI-based methods in the context of sustainable packaging. The main quantitative analysis focuses on the quality dimensions of novelty, value, and feasibility. The results indicate that GenAI models such as ChatGPT not only match, but also occasionally outperform, those generated by expert sessions in terms of generating highly novel ideas. However, this increased novelty comes with a trade-off in perceived feasibility, highlighting a critical balance that must be managed in innovation efforts. A complementary qualitative analysis provides insights into potential barriers to integrating AI into ideation at the personal and organizational levels. Depending on the innovation setting, AI-based idea stimulation may limit the creativity and motivation of experts. Therefore, the form of AI integration should be adapted to the company's innovation context and should contribute to organizational learning. On the basis of these findings, guidelines are provided on how to effectively apply, and benefit from the use of, AI as a non-human intermediary to enhance the ideation process. • We compare AI-generated ideation with traditional expert-based ideation workshops. • While AI excelled in novelty, we uncover a notable novelty-feasibility trade-off. • There were mixed reactions among employees regarding AI integration. • We offer practical recommendations on leveraging LLMs as intermediaries in ideation.