Key points are not available for this paper at this time.
Applying artificial intelligence (AI), especially natural language processing (NLP), to harness large amounts of information from patent databases, online communities, social media, or crowdsourcing platforms is becoming increasingly popular to help organizations find promising solutions. In the era of non-human innovation intermediaries, we should begin to view NLP not only as a useful technology applied in different innovation practices but also as an intermediary orchestrating valuable information. Previous research has not taken this perspective, and knowledge about its intermediation activities and functions is limited. This study reviews 167 academic articles to better understand how NLP approaches can enrich intermediation in early-stage innovation search. It identifies 18 distinctive innovation practices taking over activities like forecasting trends, illustrating technology and idea landscapes, filtering out distinctive contributions, recombining domain-specific and analogous knowledge, or matching problems with solutions. While certain NLP capabilities complement each other, the analysis shows that the choice of the most appropriate approach depends on the characteristics of the innovation practice. Innovation researchers and practitioners should rethink current roles and responsibilities in AI-based innovation processes. As seen in the recent emergence of large language models (LLMs), the rapidly evolving field offers many future research opportunities and practical benefits.
Julian Just (Tue,) studied this question.