From Information Overload to Knowledge Integration: Generative AI as an AI-Enabled Information Management Capability in Science–Industry Innovation Ecosystems
Drawing on interviews with Slovenian innovation ecosystem stakeholders, this study proposes a conceptual framework positioning Generative AI not as an autonomous intermediary but as a human-governed information management capability that addresses ecosystem fragmentation and information overload by enhancing knowledge visibility, matching, and decision support while ensuring transparency and accountability.
Original paper licensed under CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/). This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer
In the modern world of science and business, the biggest hurdle to creating new ideas is no longer a lack of information. For decades, the challenge was simply finding the right expert or the right piece of data. Today, the opposite problem has taken hold: there is too much information, scattered across too many places. Universities, companies, and government agencies are all generating vast amounts of knowledge, but this knowledge is often locked away in separate digital systems or hidden within personal networks. When a scientist wants to partner with a company, or a business seeks a new technology, they face a chaotic landscape of disconnected websites, confusing funding rules, and profiles that are hard to read. The real difficulty is not finding a needle in a haystack; it is finding the right needle when the haystack has been split into a thousand different piles, and you do not know which pile contains the needle you need.
This is the central puzzle that a team of researchers from the University of Novo mesto in Slovenia set out to solve. They wanted to understand how artificial intelligence, specifically a type called generative AI, could help untangle this mess. Generative AI is a technology capable of reading, understanding, and summarizing large amounts of text from different sources, much like a very fast, very well-read assistant. The researchers asked a simple but vital question: if we have this powerful tool, how can it actually help scientists and businesses work together? They did not assume the technology would solve everything on its own. Instead, they wanted to know what specific problems the people on the ground were facing and what kind of help they actually wanted from a computer.
To find the answers, the researchers spoke with twenty-four people who are deeply involved in the Slovenian innovation system. These participants included university researchers, business managers, and government officials who help fund and guide new projects. The team conducted in-depth interviews, asking these stakeholders to describe their daily struggles. They wanted to know where the friction lies when trying to start a collaboration. The results were clear and consistent. The people interviewed did not say they lacked information. They said the information was everywhere, but it was fragmented and overwhelming. A business manager might know that a solution exists, but finding the specific university lab that has it could require searching through ten different websites. A researcher might know a company needs help, but not know which company it is or how to contact them. The problem was not scarcity; it was the inability to connect the dots between what is available and what is needed.
The study also revealed that trust is a major factor. In the past, people relied on personal relationships to decide who to work with. If you knew someone, you knew they were reliable. In a digital world where you might not know the person, you need proof. Participants explained that they need to see a track record of past work, evidence of specific skills, and a history of successful projects before they feel safe entering a partnership. They were wary of systems that simply listed names without context. They needed to know not just who was available, but who was credible.
When the researchers asked these stakeholders what they would want from an artificial intelligence tool to help with these issues, the answers pointed toward a specific kind of support. The participants did not want a machine that would make decisions for them. They did not want an automated system that would pick their partners or write their grant applications without their input. Instead, they envisioned the AI as a powerful assistant that could organize the chaos. They described five main ways this tool could help. First, it could act as a smart matchmaker, connecting a company's specific needs with a researcher's specific skills, looking beyond simple job titles to find true compatibility. Second, it could help navigate the complex world of funding, filtering through thousands of grant opportunities to find the few that actually fit a project's goals. Third, it could make expertise visible, creating clear profiles that show not just what a person does, but what they have actually achieved. Fourth, it could help with the heavy paperwork of writing proposals, summarizing rules and drafting documents so people could focus on the ideas. Finally, and perhaps most importantly, it could support decision-making by gathering all the relevant facts and presenting them clearly, while leaving the final choice to the human.
The researchers found that for this technology to be trusted, it had to be transparent. The people interviewed insisted that they needed to understand why the AI suggested a particular partner or funding source. They wanted to see the evidence behind the recommendation. They also demanded that their data be kept secure and that a human always remained in charge of the final decision. The AI was seen as a way to handle the heavy lifting of information management, but the human was seen as the one who must judge the value of that information and build the relationship.
This study suggests that the future of science and industry collaboration does not lie in replacing human intermediaries with robots. Instead, it lies in using artificial intelligence to handle the information overload that currently paralyzes the system. By organizing fragmented data, verifying credentials, and filtering out the noise, these tools can make the vast ecosystem of knowledge accessible and usable. The technology does not create the trust or the partnership itself; it clears the path so that humans can do what they do best: judge, connect, and collaborate. The researchers conclude that if these systems are built with transparency and human oversight at their core, they can transform the current state of confusion into a streamlined environment where new ideas can finally take root.
Drowning in papers in your field?
Get daily digests of the most novel papers matching your research keywords — with technical summaries, in your language.