Microsoft is creating a self-adaptive AI system, a steerable virtual scientist, that can think, plan, and solve fundamental scientific problems in areas like material design and pharmaceuticals.
This AI, built using a CLIO method, lets scientists guide its reasoning, step in when needed, and understand how it reaches conclusions.
It even improved GPT-4.1’s performance in biology and medicine from 8.6% to 22.4% without additional training.
Unlike typical AI tools with fixed reasoning, CLIO works in real time, adjusting its thinking using internal feedback loops.
Researchers can set thresholds for the AI’s confidence, steer its decision-making, and see its reasoning steps clearly.
This transparent and controllable approach brings much-needed trust and insight to high-stakes scientific work.
You may also want to check out some of our other recent updates.
Subscribe to Vavoza Insider to access the latest business and marketing insights, news, and trends daily! 🗞️
