This three-hour workshop aims to foster a community-wide discussion on the impact of the shifting relationship between academic and industry AI research on iSchool AI researchers. Traditionally, academic and industry research serves different purposes, with academia focusing on theoretical and foundational research, and industry emphasizes on practical applications. In recent years, the growing role of industry is changing the relationship between academic and industry research on AI. Compared to the industry, academic AI research faces significant challenges, especially limited computing resource, access to large data sets, and a shortage of talent. These challenges undermine academia's role as checks and balance toward responsible and trustworthy AI.
iSchools are uniquely positioned in AI research landscape with their interdisciplinary nature and a human-centered AI research agenda. This workshop invites iSchool AI researchers and research administrators to join the discussion, sharing our concerns and brainstorming new ideas for navigating and influencing this shift. For example, what can we learn from this shift? How does it affect our decisions to choose research topics and acquire research resources? How do we position ourselves in this shifting relationship? What are the main concerns in our research community? How do we as a community envision the near- and long-term impact of this shift on research?
This workshop includes expert panels, lightning talks, group discussions. We welcome iConference attendees to join the workshop, sharing our concerns and brainstorming new ideas for navigating and influencing this shift.
Main points from the panel, presentations, and group discussions will be collected and summarized into a white paper and shared with the workshop participants.
We will also collect email addresses of participants for continuing discussion after the workshop, such as exploring the possibility of developing a special issue for JASIST or similar journals.
Traditionally, academic and industry research serves different purposes, with academia focusing on theoretical and foundational research, and industry emphasizes on practical applications. Collaboration between them has helped translate academic research to real-world products and services.
In recent years, the growing role of industry is changing the relationship between academic and industry research on AI. For instance, in the field of natural language processing (NLP), industry now publishes significantly more top-cited papers, especially on foundation models and relevant computational methodology (Movva et al., 2024). A few companies account for most of the industry publications on NLP, and they provide funding to academic researchers through grants and internships (Abdalla et al., 2023). Furthermore, the industry-academic collaborations tend to center around topics aligned with industry interests, raising concerns on academic independence in choosing research topics. With industry leading on AI model development, there is increased risk that academic researchers may be left to use industry models for applied tasks only. In 2023, big tech companies had reduced their research publications (Movva et al., 2024), adding concerns over increased secrecy of industry research (Bommasani et al., 2023).
Compared to the industry, academic AI research faces significant challenges, especially limited computing resource, access to large data sets, and a shortage of talent. Within academia, the ability to independently develop and train large AI models is restricted to a small number of leading universities, raising concerns on inequality of academic AI research (Swaak, 2024). Academia should play an important role as checks and balance for responsible AI research. The imbalance between academic and industry research increases the risk of AI misuse. For example, although academic research has produced knowledge on responsible AI, large AI companies are reluctant to apply it (Owen, 2024).
For inquiries, please contact Bei Yu (byu@syr.edu).