This week saw major developments in AI agents transforming scientific research. At the UC Berkeley Agentic AI Summit in the US, experts discussed how self-directed AI systems could accelerate discoveries by planning experiments and analyzing data independently.

IBM researchers highlighted new agentic AI capabilities like better reasoning and tool usage, which are helping scientists process complex datasets faster than ever before. A Forrester report predicted these systems will soon become essential for competitive research by handling tasks that once required human teams.

In healthcare, Causaly’s Agentic AI platform demonstrated how AI can generate testable hypotheses for diseases by finding hidden patterns in millions of research papers – cutting discovery time for some projects by over 60%.