What is robotics made of? The interdisciplinary politics of robotics research

Robotics is increasingly integrated into various aspects of society, presented as a solution to grand challenges like road safety, aging populations, and climate change. However, the development of robotics is not apolitical; it's a site of political contestation and choices. This paper focuses on the politics of interdisciplinary robotics research, specifically examining how it's co-produced within research institutions and wider society. The research questions explore the politics of interdisciplinary robotics research, what this reveals about the composition of robotics, and how to structurally prepare research projects for interdisciplinarity to align with public interest. The paper analyzes politics through three lenses: cartographically (disciplines involved and integration modes), discursively/institutionally (interpretations of robotics' potential), and capably (everyday choices and practices of researchers). The study centers on the Bristol Robotics Laboratory (BRL) to investigate these aspects.

Existing research on interdisciplinarity highlights its societal concern-driven and politically influenced nature. The paper draws on co-production theory, emphasizing the mutual implication of science and society in robotics development. It also reviews different conceptualizations of interdisciplinarity, including Kelly's 'wide' and 'narrow' interdisciplinarity, and the importance of considering both disciplinary relationships and research practices. The authors cite previous work emphasizing the shift from solely measuring disciplines to investigating lived experiences and practices in understanding interdisciplinary collaborations, highlighting the importance of capabilities beyond cognitive abilities for effective interdisciplinary research. The role of accountability, innovation, and ontological change as drivers for interdisciplinarity is also discussed, acknowledging the limitations of reducing interdisciplinarity to these factors. The paper advocates for mixed-methods approaches to analyze the multifaceted nature of interdisciplinarity.

The study employed a novel mixed-methods framework to conduct an in-depth case study of the Bristol Robotics Laboratory (BRL). Bibliometric analysis of peer-reviewed publications (Scopus and Web of Science) from 2004-2020 visualized interdisciplinarity at an institutional scale, identifying prominent disciplines and keywords. Content analysis of publicly available documents from 63 recent BRL projects provided insights into project-level collaborations and disciplinary integration, examining how interdisciplinarity is conceptualized and enacted. Three embedded case studies, based on observational research and document analysis, explored researchers' experiences and valued capabilities in specific research areas: self-driving vehicles, bioenergy, and assisted living robotics. The researchers used VOSviewer for visualization of bibliometric data, categorizing Web of Science disciplines into machine software, hardware, environment, medicine, and human factors/society clusters. Content analysis used a framework charting collaborations (academic and non-academic), categorizing interdisciplinarity (wide, medium, narrow), and analyzing project framings (societal challenges, industry needs, theoretical). The authors acknowledge their positionality as early-career researchers, offering a critical yet constructive perspective.

Bibliometric analysis revealed a dominance of engineering and computer science disciplines in BRL publications, with a less prominent role for social sciences and humanities. Keyword analysis showed frequent use of 'human' in various contexts (interaction, experimental subjects, beneficiaries), but limited representation of social sciences in publications. Content analysis of 63 projects indicated computer science, mechanical engineering, and psychology as most common disciplines. Projects were categorized as 'wide', 'medium', or 'narrow' interdisciplinarity based on disciplinary breadth. Framing analysis showed a majority of projects focused on societal challenges, often involving non-academic partners. However, some projects were solely industry-driven or theoretically focused. Case studies highlighted different logics underpinning interdisciplinary research: bioenergy (experimentation, tackling poverty), autonomous vehicles (testing innovation, market growth), and assisted living (innovation as crisis response, market creation, accountability). Each case revealed specific capabilities developed by researchers, such as aligning funder needs with lab narratives, building diverse stakeholder networks, but also a lack of capabilities to fully steer innovation and challenge underlying assumptions about technological progress. The analysis noted disparities in stakeholder involvement, with significant industrial partnerships but limited inclusion of voluntary and community sector organizations. The findings underscore the plurality of interdisciplinary practices within BRL and the varied ways societal needs are addressed.

The findings demonstrate that interdisciplinarity within BRL is not uniform, shaped by funding strategies, collaborations, and researchers' choices. The study reveals how different actors (funders, researchers, industry partners, end-users) possess varying capabilities to influence research directions and highlight power dynamics in knowledge creation. While large consortia might involve diverse stakeholders, they may not necessarily disrupt power structures. The authors emphasize the importance of carefully assembling interdisciplinary teams to ensure diverse voices and challenge existing power arrangements. The study's focus on BRL illustrates how different logics and infrastructures shape interdisciplinary research, and how this interacts with real-world needs and funders' expectations. This highlights how interdisciplinary activities adapt to fit perceptions of the world, negotiated between researchers and funders. The study concludes by emphasizing the importance of cultivating capabilities to steer innovation, challenge assumptions, and meaningfully involve societal partners to address grand challenges.

This research offers a nuanced understanding of interdisciplinary politics within robotics research, demonstrating the diverse approaches, capabilities, and power dynamics at play. The study highlights the need for reflective practices within robotics research to ensure inclusive and responsible innovation. Future research should focus on strengthening capabilities to steer innovation, challenge assumptions, and meaningfully incorporate societal partners and critical social sciences, thereby ensuring that robotics research reflects and serves the values and interests of society.

The study's focus on a single robotics laboratory (BRL) limits the generalizability of findings. Data limitations, including reliance on publicly available documents and potential biases in databases, might have influenced the results. The research team's positionality as early-career researchers also needs to be considered, although this also provided a critical yet constructive perspective. Further research is needed to explore the generalizability of the findings to other robotics laboratories and research contexts.