Governing the Use of AI in Scientific Research
Thematic track at Eu-SPRI 2025:
https://euspri2025.de/index.php/callforpapers/
convenors: Laurens Hessels, Anne van Doore, Kieron Flanagan
deadline for abstract submission 31 January 2025
Machine learning techniques have accelerated the development of artificial
intelligence (AI) in recent years, leading to a proliferation of new
applications. AI technologies are transforming many domains of our society and
economy, such as mobility, entertainment and public administration.
The rapid development of AI is also generating many new ways of using AI in
scientific research. Until recently, the use of AI was concentrated in data
analysis and programming, but the introduction of user-friendly generative AI
tools such as ChatGPT has opened up new applications in other parts of the
research process, such as literature review and text writing. We can safely
assume that AI will have disruptive effects on the science system in the
coming years.
AI brings a number of great potential benefits, in particular increasing the
productivity of research (OECD, 2023). Large language models can accelerate
the research process by generating summaries of notes or even carry out
complete literature scans. AI can also facilitate efficient data mining,
qualitative analysis or assist in writing grant applications. Generative AI
could also help reduce geographic inequality between researchers by improving
articles in terms of writing quality and language errors.
However, there are also serious risks involved. The risk of discrimination is
significant, particularly because AI models are trained on a sometimes unjust
reality based on stereotypes and biases (Nicoletti & Bass, 2023). Critics have
also raised concerns about the incompleteness of summaries (Barak-Corren et
al., 2024), the numerous errors that occur, and thus the unreliability and
inconsistency of systematic reviews (Gwon et al., 2024), as well as the loss
of control over copyright (Appel et al., 2023). Additionally, directly copying
or paraphrasing from a chatbot raises questions about plagiarism (Ciaccio,
2023). The use of machine learning algorithms also tends to make the research
process less explainable, undermining the principle of replicability (Ball,
2023). Finally, the use of AI consumes so much energy, that it is highly
uncertain whether its carbon footprint is proportion to the additional
benefits gained by AI.
Altogether, it is clear that the use of AI in science deserves scrutiny. In
order to enable researchers to make use of AI in a responsible way,
professional guidelines or public policies will probably be necessary. Some
scientific publishers and funding agencies have formulated (tentative)
guidelines, largely focused on AI in the writing process, but in many
situations researchers are left without any guidance.
This raises a number of questions with science policy and governance
implications, such as:
- What does AI mean for the nature of scientific knowledge and for the
identity of researchers (across different fields)?
- How is AI changing the social processes by which the scientific community
determines which knowledge claims are valid and which methods reliable? Will
the large-scale use of AI require new ways to ensure scientific integrity and
reliability?
- What competencies do the scientists of the future need, given the rise of
AI?
- What are the implications of the rise of AI for academic career paths and
job markets? Will we need fewer scientists in future?
- What new opportunities does AI offer for open science and transdisciplinary
collaboration? And where does it pose a challenge to them?
- How can research security be ensured when scientists make extensive use of
AI, which partly depends on services and tools from foreign private
providers?
In this track we aim to bring together researchers from science policy
studies, STS, philosophy of science and related fields interested in the
effects of AI on research practices in academic institutions and other public
sector organizations, and in the potential of public policies and
organizational policies for facilitating and ensuring the responsible use of
AI.
EASST's Eurograd mailing list
Eurograd (at) lists.easst.net
Unsubscribe or edit subscription options: http://lists.easst.net/listinfo.cgi/eurograd-easst.net
Meet us via https://twitter.com/STSeasst
Report abuses of this list to Eurograd-owner@lists.easst.net
