Last week, I attended the AI for Science Forum, a gathering of incredible minds from across disciplines, each sharing perspectives on how AI is transforming research and impacting society. Organised by Google DeepMind and The Royal Society, the event brought together invited guests from across all segments of the research community to share their experiences and expertise, while also giving opportunities for attendees to meaningfully discuss how we can best wrangle this novel technology to increase the impact and reach of research with the resources currently at our disposal.
The Power of AI in Research
Fresh-faced after the previous evening’s reception at The Royal Society, James Manyika, Senior Vice President at Google Alphabet, set the tone for the conference with a powerful opening session on the tangible benefits of AI. From AlphaFold’s protein-folding revolution to AI’s role in flood forecasting in Bangladesh which has already impacted millions of people, his talk was a staggering reminder that AI is already being used in various innovative and impactful ways beyond those we are more familiar with. He also discussed AI-enabled solutions through the lens of a public health research focus, reminding attendees of advancements including diabetic retinopathy detection in low-resource settings and the groundbreaking atmospheric simulations helping better prepare agricultural workers with advanced warning of what interventions may be required. This reminded me of our recent TL;DR Shorts episode with Dr Danny Hillis of Applied Invention who talked about the potential impact that automated research could have in helping us help non-traditional researchers. But James’s outlook for AI wasn’t all rosy as he underscored the limitations of this tool and emphasised the need for responsible approaches and equitable access to AI-powered tools, echoing his colleague Dr Astro Teller’s thoughts on this.
CRISPR Meets AI
Nobel Prize-winning Chemist Professor Jennifer Doudna, Professor of Biochemistry, Biophysics and Structural Biology at UC Berkeley, and James Manyika picked up on this theme as they explored the synergy between CRISPR and AI. They discussed how CRISPR’s one-and-done gene therapies are accelerating in application thanks to AI’s ability to identify the genetic changes that drive outcomes, and democratising treatment options by providing a range of more affordable therapeutic options. Beyond healthcare, they also chatted about the impressive potential impact that AI will have on climate-related research, from drought-resistant crops to better carbon storage systems. What resonated most with me was their call to reduce the barriers – financial, technical, and geographic – to accessing the outcomes of this technology, making it truly global, and reflecting the recent thought we shared from Professor Lord Martin Rees.
Collaborating Across Disciplines
The focus shifted to the future of collaboration in a panel led by Eric Topol, author and Executive Vice President of Scripps Research, and featuring Fiona Marshall, President of Biomedical Research at Novartis, Alison Noble, Oxford University Technikos Professor of Biomedical Engineering, and Vice President & Foreign Secretary, The Royal Society, and Pushmeet Kohli, Vice President of Science, Google DeepMind. From AI revealing 2.2 million new stable inorganic crystals for potential use in everything from energy to electronics, to revolutionising biomedical imaging through natural image processing, the discussion highlighted how AI forces us to rethink and redefine what collaboration and trust look like. Alison’s comment on the importance of training scientists to understand AI’s errors stood out for me. The panel emphasised the shifting hierarchies and power dynamics in research, with data scientists increasingly leading labs – a significant cultural change, given that they were often seen as collaborators and service providers who were rarely even named on papers that couldn’t have been published without their expertise.
Climate, Complexity, and Community
Thomas Friedman gave an evocative talk on “climate weirding”, highlighting how we’re hitting many tipping points all at once – AI’s massive breakthroughs, climate change chaos, and global instability – while also linking AI capabilities to potentially help us with carbon emissions and societal disorder. His call for politics to embrace science felt especially urgent amid discussions of mass migration and global conflict. His optimism about how AI can solve big problems, like making farming more efficient or cutting healthcare costs, came with a warning that we also need solid ethics and politics to go hand in hand with these developments to keep things on track, something that our recent Speaker Series guest Professor Jenny Reardon touched upon.
The next panel discussion on building research infrastructure echoed these sentiments. Chaired by Paul Hofheinz, President and Co-Founder, Lisbon Council, and featuring Asmeret Asefaw Berhe, Professor of Soil Biogeochemistry and Falasco Chair in Earth Sciences, University of California, Merced, Bosun Tijani, Minister of Communications, Innovation & Digital Economy, The Federal Republic of Nigeria, and Fabian J. Theis, Director of the Institute of Computational Biology and Professor at TUM Mathematics & Life Sciences, both Paul and Asmeret stressed the need for equity and inclusion to be at the forefront of people’s agendas as they develop solutions using AI, to ensure that advancements don’t deepen the digital divide. Bosun Tijani’s discussion of Africa’s talent acceleration programs was inspiring – and proof that we can nurture talent globally if we commit to the cause. We recently heard from Joy Owango about how important it is to build infrastructure that ensures the persistence and visibility of research contributions from all across the globe, and how impactful this has already been in Africa and other parts of the Global South.
This theme continued as Lila Ibrahim, Chief Operating Officer at Google DeepMind chaired a conversation with Dame Angela McLean, UK Government Chief Scientific Adviser, Ilan Gur, CEO, Advanced Research and Invention Agency (ARIA), and Sir Paul Nurse, Director of the Francis Crick Institute, Nobel laureate, and returning President of the Royal Society chatted about collaborating for impact. The panellists discussed the importance of thinking big and including diverse perspectives through better community engagement in science, something that X’s Dr Astro Teller talked about in a previous TL;DR Shorts episode. Dame Angela talked about how the government needs to aim higher, pushing for more thoughtful use of AI and predictive models, while Sir Paul stressed the need for mixing disciplines to boost innovation. Ilan shared his excitement about creating spaces where scientists from different fields can cross paths, sparking unexpected ideas. Recorded at Sci Foo, a perfect example of a catalyst for collaboration, Dr Etosha Cave echoes this sentiment and the need for interdisciplinarity for innovation. The panel also discussed building trust, with Angela and Ilan both emphasising the importance of transparency in science and technology. All panellists highlighted the role of public engagement in encouraging people to engage with and trust these cutting-edge advances.
Public Engagement and Trust
The conference ended with a final discussion featuring recent Nobel Prize winners in Chemistry, Sir Demis Hassabis and John Jumper, as well as former winners Professor Jennifer Doudna, and Sir Paul Nurse. Their reflections on public engagement were poignant: how do we bridge the gap between experts and the public? Sir Paul’s call for deliberate public dialogue reminded me how crucial it is to address fears and misconceptions about AI before they grow into barriers. However, one issue that continually cropped up, and one I may have mentioned once or a million times in the past, is that as it stands, the framework within which we reward research success does not make space for valued and impactful public engagement, or even innovation and entrepreneurship. Mariette DiChristina had a few thoughts on this, and we’ll be hearing more from her in 2025 about the value of effective communication of, and engagement with research in the age of open research, research integrity, and novel technology such as AI.
Some Key Takeaways
- AI isn’t just transforming research; it’s reshaping the cultures around it. We’re seeing shifts in leadership, collaboration, and the ethical frameworks underpinning research.
- Accessibility remains a challenge. Whether it’s CRISPR or AI infrastructure, we need to ensure the benefits reach everyone, not just the privileged few.
- Collaboration is more vital than ever. From breaking disciplinary silos to engaging the public, success hinges on our ability to connect diverse voices.
In a world increasingly shaped by AI, this conference left me both hopeful and reflective. Science thrives when it’s inclusive, transparent, and collaborative – and AI could give us a chance to embrace those ideals like never before, provided we build research methods and applications in thoughtful, considerate, trustworthy and community-minded ways. My teammates recently authored a report that, in true Digital Science style, was informed by reflections from our own research community. The report looks at the changing research landscape in the age of AI and echoes the many challenges and opportunities discussed at this conference. AI is an exciting development that is already changing the way we do research. However, we must hold each other accountable to ensure that its development and application are open to all.
With thanks to Google DeepMind and The Royal Society for hosting this event. You can watch all sessions on Google DeepMind’s YouTube channel.
