How will AI influence preterm birth research?
Artificial intelligence, or AI, is becoming an increasing part of everyday life and conversation. Chief Scientific Advisor Michael Taggart, together with Borne-funded researcher Pei Lai, considers possible implications for science and pregnancy and preterm birth research
AI, an acronym for Artificial Intelligence, is a term that generates confusion, concern or unbridled enthusiasm depending on one’s perspective. It can be viewed as a catch-all phrase to describe current activity, and future possibilities, in the digital workspace that harnesses decades of incremental increases in the capabilities of language (coding), software (the operating features of computer models), hardware (parallel processing chips) and computing (so-called machine learning). It is the integration of these processes that has now produced possibilities for advancing research that has captured the imagination of scientists, engineers, digital tech companies and governments. The UK government, for example, has established an ‘Incubator for Artificial Intelligence’ to oversee the testing and adoption of AI processes across government and the public sector, has instigated two public consultations on the implications of AI for issues of copyright and data security, and has appointed a Minister of State for AI.
Borne-funded researcher Dr Pei Lai (Imperial) attended a recent Nobel Prize Dialogue event, hosted by the Royal Society in London, entitled ‘The Future of Science with AI’. The discussants included Paul Nurse (Fellow of the Royal Society, former Director of The Crick Institute, and Nobel Laureate in Physiology or Medicine) and Demis Hassabis (Fellow of the Royal Society, co-founder of Google DeepMind, and Nobel Laureate in Chemistry). You can view a recording of the discussion here: The Future of Science With AI | Nobel Prize Dialogue London 2026.
Demis Hassabis commented on the dramatic role that AI has played in his own research field. His DeepMind research team developed AlphaFold as an AI system to digitally predict the three-dimensional structures of proteins from their amino acid sequences, which was made possible using data from existing human genome/proteome databases. DeepMind shared over 200 million protein structure predictions in an open-access format; these are free to use for researchers worldwide for analysis, validation and to inform their own studies. This has not only reduced the time to discovery but also expanded future scientific research possibilities.
Similarly, Paul Nurse highlighted that AI tools have been useful in reducing the time taken to perform specific research-related tasks, such as literature searches and analyses of large datasets, with the attendant benefit of assisting the generation of new hypotheses. He emphasised though, it was imperative that the critical thinking behind every step of an experiment undertaken to study biology remained with the human rather than a computer. This especially holds when one realises many advances in science have arisen from unconventional approaches and serendipity.
Dr Lai comments on the meeting: “As a ‘wet lab’ scientist exploring how I can utilise AI for my pregnancy and preterm birth-related research, it was a validating experience to have related to, and agreed with, several points raised.. Most notably, we need to (i) view AI as an assistant that deals with some (not all) time-intensive repetitive tasks in a way that liberates space for the deep thinking needed to solve complex scientific problems; (ii) move away from a culture that rewards simply being first to use the newest technology for generating big data and instead focus on ensuring the quality of these outputs yield new biological knowledge; (iii) thoughtfully factor in the complex and dynamic nature that exists within cells, which will be essential for setting realistic goals and timescales for what AI can achieve to help improve human health”.
What are the implications of AI for pregnancy and preterm birth research?
There are two main points that come to mind.
First, there are opportunities to be garnered from other scientific fields. For example, AI is advancing predictions of how to design and synthesise peptide (protein)-based antibacterial agents and this could be useful for preventing/treating cervical infections (a contributor to spontaneous preterm birth). Also, AI improvements in image analysis may improve the accuracy, and speed, of extracting key features from ultrasound scans; this could potentially assist in antenatal screening for pregnancy complications including defining more clearly women at risk of preterm birth.
Second, for pregnancy and preterm birth research per se the application of AI is in its infancy. This, in large part, is because there is a relative paucity of relevant data to inform AI model utilisation targeted to key questions, certainly when compared to data-rich biology fields that have dominated so far e.g. cancer, neurology, bacteriology and genomics. If anything, the immediacy of AI emphasises the importance of Borne’s mission to fund more fundamental scientific, and applied clinical, research to understand pregnancy and preterm birth and furnish AI tools with relevant data. That said, there are likely to already be opportunities for the use of AI models to facilitate the analysis of current, often disparate, pregnancy and preterm birth research datasets. For example, AI models could assist to collate in one space, and make user-friendly for further integrated interrogation, large molecular datasets on cervical bacterial colonies, or those focussed on mechanisms of uterine muscle activation for labour, from many different studies worldwide that could serve to answer important questions about preterm birth.
Dr Lai adds: “I think it’s important to also note that the research for which Demi Hassabis received the Noble Prize in Chemistry was only possible due to private sector investment from Google. If we are to utilise AI tools in the most innovative way for answering complex questions around pregnancy and preterm birth, it’s likely that research funders will have to consider connecting with private research institutions, particularly those with the infrastructure to invest in bold ideas, in addition to academic experts within universities”.
Borne recognises the on-going importance of these considerations and the need to keep abreast of latest research developments. As such, we look forward to the forthcoming Borne Collaborative meeting in Glasgow, March 2027 when leading international scientists and clinicians will debate the topic ‘How will AI influence preterm birth research?’
