Artificial intelligence (AI) significantly expedited the development and deployment of Covid-19 vaccines, as scientists harnessed machine learning to analyze extensive datasets, including journal articles and clinical trials. This enabled a swift prioritization of research pathways, circumventing traditional obstacles in vaccine development while enhancing the accuracy of antigen selection.
In the United States, the legal framework supports the utilization of such data by commercial organizations and academic institutions alike, as evidenced by the work of Pfizer-BioNTech and Moderna in creating mRNA vaccines. In contrast, UK copyright law presents challenges for AstraZeneca, which collaborated on the Oxford-AstraZeneca viral vector vaccine. Under current regulations, AstraZeneca researchers would face restrictions in using AI to automate data extraction from potentially relevant papers, as they would have to manually sift through extensive materials, armed only with a pen and paper.
For-profit organizations like AstraZeneca would need permissions from each publisher to deploy AI for data mining, which incurs copyright liabilities due to the nature of machine learning. Additionally, obtaining permission from social media platforms would be necessary for using AI to analyze vaccine skepticism or identify areas with high infection rates to optimize vaccine distribution. This cumbersome process poses a significant hurdle, as licensing experts note the impracticality of seeking approvals for every piece of data utilized in AI training.
The UK government’s AI for Science Strategy underscores three essential pillars: people and skills, technical infrastructure, and data. However, the emphasis is heavily placed on technical infrastructure while attempts to address data issues often avoid the complexities of copyright, focusing instead on the creation of new datasets.
The recently announced US-UK “Tech Prosperity Deal,” which aims to attract US investments in the UK’s AI sector, raises questions about the types of data that will be processed there. Effective utilization of Big Data is contingent on a legal framework that facilitates its flow; without conducive data laws, even the most advanced infrastructure risks being underutilized.
This paradox highlights the absurdity of current copyright laws, which aim to protect original expression yet inadvertently stifle access to vital information. Facts, trends, and relationships are meant to remain in the public domain under international copyright treaties. However, as these data points exist within copyrighted materials like articles and books, UK law necessitates permissions for data mining—despite the absence of intent to replicate original works.
The implications of such restrictions are dire for the speed and efficacy of research, particularly given the UK’s focus on commercializing scientific breakthroughs as part of its industrial strategy. The government’s year-old consultation on AI and copyright frequently references large language models (LLMs), yet the landscape has largely been dominated by American and Chinese players, leaving the UK lagging behind in the race for AI innovation.
Despite these challenges, the UK can carve out a niche within AI markets by focusing on sectors where it already excels, such as biomedicine and finance. The emergence of DeepSeek presents a potential turning point for UK startups, universities, and SMEs. This open-weight LLM, despite concerns over its Chinese origins, allows developers to access its trained parameters freely, fostering innovation.
For a government keen on revitalizing the economy, investing in open weights and modernizing data laws could prove strategic. A robust legal framework around data would enable the UK to compete on equal footing with the US and East Asia. Observing the successful approaches taken by Japan, South Korea, and Singapore— which amended their copyright laws nearly a decade ago to foster AI and data-driven innovation—could offer valuable insights for UK policymakers.
To establish a comprehensive science-focused strategy, it is crucial for university and research leaders to advocate for policy changes that reflect the broader implications of AI and copyright. The recent government consultations have largely overlooked the significance of these issues for the scientific community, focusing primarily on creative industries and Big Tech.
As a research-intensive economy, the UK’s prosperity hinges on fostering a vibrant scientific community. Policymakers must recognize that investing in physical infrastructure alone is insufficient; a flexible and modern approach to data and copyright laws is essential to unlock the full potential of AI-driven research and innovation.
See also
Uolo Secures $7M Funding to Enhance AI Learning for 1.1M Students in K–12 Schools
Instructure Launches IgniteAI with Key Updates for Canvas, Enhancing Data Privacy and Educator Support
White House Hosts Third AI Education Task Force Meeting to Boost K-12 AI Literacy
OpenAI Launches GPT-5.2, Achieving 92.4% on PhD-Level Science Benchmarks
UK Government Partners with Google DeepMind to Launch Automated Research Lab in 2024
