Berlin | Claude Conversation x Foresight Institute — Can AI help us solve science's hardest problems?
AI systems can now read the literature, write and run analysis code, and propose experiments. Some research groups already run agents in a loop against their own open problems. Others see AI mostly as a source of confident mistakes, or as a threat to how science is done and who gets to do it.
This is an evening conversation about one hard question:
What would it take for AI to speed up the science that matters most, and what is left for us to do?
We will bring together working scientists from the life sciences, medicine, chemistry and materials, physics, mathematics, nanotechnology, and climate and earth sciences with the engineers who build AI agents. This is not a lecture and not a product demo. We will spend most of the evening in small groups around four questions:
How will this change our field? Where is AI already useful in your research, where does it fail, and what worries you?
What do we need to address this? Data, tools, skills, verification, trust: what is missing between a capable model and a result you would publish?
What can we do now? Which open problems in your work are ready for an AI research loop today?
Where do we begin? Together we pick the concrete problems we will build on.
The evening ends with a shortlist of real open research problems. These become the brief for the follow-up Claude Impact Lab, a full-day build where scientists and AI engineers turn them into working, open-source research loops.
The series is co-hosted by the Foresight Institute Berlin AI Node. It continues Foresight's Existential Hope work on how advanced technology can go well for people, starting from the problems scientists actually face.
Who it's for: researchers from any field and at any career stage who want to shape how AI enters their field, and AI engineers who want their work to matter outside software. No technical background needed.
Bring: a problem from your own research that you wish someone or something would take off your hands.
Free admission. Participants are selected by application. Hosted by Kazik Pogoda, AI researcher and volunteer Claude Ambassador for Science working with the Foresight Institute on epistemic AI tools for science, and Keith Patarroyo, Foresight Institute fellow at the University of Glasgow. Together they presented the AI4Science Buildathon concept that won a prize at Foresight's 2026 Secure & Sovereign AI Workshop, and this series is its first edition.