

“Scientific Discovery in the Age of Artificial Intelligence” by Jessica Rumbelow
Jul 1, 2025
Jessica Rumbelow, an author focusing on AI's potential in scientific discovery, shares her insights on leveraging artificial intelligence for accelerated research. She discusses the transformative power of machine learning in medicine, especially its role in understanding aging and drug resistance. Rumbelow also highlights tools like knowledge graphs and robotic labs, emphasizing the collaboration between AI and researchers. With promises of superhuman medicine and radical life extension, she explores the exciting future of humanity's quest to understand the universe.
AI Snips
Chapters
Books
Transcript
Episode notes
Core Scientific Process Challenges
- Scientific progress relies on iterative hypothesis testing and publishing results to build collective knowledge.
- Many AI tools currently assist parts of this process but face challenges like imperfect data and non-replicability.
Limits of AI Literature Reviews
- AI literature review tools use semantic search and LLMs to find and summarize papers.
- Coverage is incomplete and summaries have 10-20% error, plus unreliable peer review worsens trustworthiness.
Knowledge Graph Success Story
- Benevolent AI used knowledge graphs to find 378 candidate drugs, identifying baricitinib to reduce COVID mortality.
- Nonetheless, only a few candidates proved effective due to extraction errors and unreliable literature.