AI in Science

Why AI in Science?

Scientific discovery is entering a new phase. AI systems are beginning to generate hypotheses, design experiments, and, in some cases, execute them. Over time, this could compress decades of progress into years. Realising that potential depends on more than access to models. It requires new infrastructure, new funding mechanisms, and new ways of organising research. Without these, AI risks being applied incrementally rather than transforming how science is done.

ARIA’s role is to help create the conditions where this transformation can happen and to apply it to the hardest scientific problems.

Leadership

Ant Rowstron joined ARIA as Chief Technology Officer, with a remit spanning the agency’s technical direction. It quickly became clear that a new opportunity was emerging: advances in AI were beginning to have a direct and profound impact on how science could be conducted. Rather than sitting alongside ARIA’s work, this shift cut across all of it, affecting how programmes are designed, how research is executed, and how quickly discovery can happen.

In response, this work evolved into a dedicated role. In 2026, Antony moved into the role of Chief AI Scientist Officer to lead ARIA’s strategy on AI in Science. His focus is to push the frontier of AI for discovery, while enabling ARIA’s Programme Directors and Creators to move faster using emerging AI capabilities. This includes both applying today’s most advanced systems and exploring what comes next.

Our approach

We are taking an intentionally experimental approach, organising our work around three interconnected pillars: Capital, Capabilities, and Communities. 

We are deploying around £10M per year to catalyse AI-enabled science, supporting everything from early exploration to full-scale initiatives in areas like automated experimentation, AI safety, and next-generation research infrastructure. In parallel, we are building and accessing the capabilities needed to accelerate discovery, from advanced AI reasoning systems and high-throughput laboratories to emerging tools that can support programme design, knowledge generation, and creator engagement. Alongside this, we are working to shape a coherent and ambitious AI in Science ecosystem – developing thought leadership, defining key challenges, and connecting communities across disciplines such as AI, life sciences, robotics, and materials science. We will adapt this approach as we learn, doubling down on what works and changing what doesn’t, with the aim of ensuring ARIA acts as a proactive leader in a rapidly evolving field.