ORNL Advances Autonomous Labs with AI-Driven Decision-Making

Oak Ridge National Laboratory is pushing the boundaries of autonomous scientific research through artificial intelligence systems capable of making experimental decisions with minimal human intervention, according to Rob Moore, a leader in the lab's self-driving laboratory initiatives.

Moore, who directs ORNL's INTERSECT initiative and the new Labs of the Future program, outlined the distinction between automation and true autonomy in a recent interview first reported by Oak Ridge National Laboratory. The difference hinges on decision-making capability.

Why it matters

Autonomous laboratories could dramatically accelerate solutions to complex scientific challenges that have remained unsolved for decades. By offloading cognitive tasks to AI systems, researchers can focus on higher-level problem-solving while machines handle iterative experimental decisions — potentially transforming the pace of materials discovery, quantum computing research, and other fields critical to national competitiveness.

From automation to autonomous operation

Moore draws a clear line between automated and autonomous systems. An automated laboratory can execute repetitive tasks without human intervention, but it lacks the capacity to respond to unexpected findings. ORNL's 4-D scanning transmission electron microscopy system exemplifies this: it uses neural networks to identify and classify atomic defects in images, but stops there.

"Full autonomy is seeing something interesting and being willing to pull the thread to dive deeper," Moore explained. "Having decision-making in there to drive the next set of experiments, this is how we can think about the difference between autonomy and automation."

True autonomous operation requires AI agents that can formulate hypotheses, steer experiments based on real-time results, and determine next steps — with humans providing oversight rather than direct control.

The reliability challenge

The rapid development of large language models has opened new possibilities for scientific applications, but Moore emphasizes that accuracy remains paramount. AI systems can "hallucinate" — generating plausible-sounding but incorrect information — which poses unique risks in scientific contexts.

"Science is a little different. We can't produce bad information for society," Moore said. "The information we put out there, we have to do our due diligence to ensure that it is accurate, reliable and reproducible."

Despite these challenges, AI excels at identifying correlations in complex datasets far faster than human researchers, making it valuable for hypothesis generation and experimental steering.

Building toward Genesis Mission goals

Moore's work supports the Department of Energy's Genesis Mission, a national initiative to create what officials describe as the world's most powerful scientific platform for accelerating discovery science, strengthening national security, and driving energy innovation.

The Labs of the Future initiative builds on ORNL's INTERSECT program, which developed a scalable ecosystem for interdisciplinary self-driving research processes. Moore joined ORNL in 2019 after serving as a U.S. Navy submarine officer, bringing expertise in quantum materials synthesis and characterization.

Moore noted that when large language models first emerged, their potential impact on scientific research wasn't immediately obvious. The speed of their development surprised researchers and revealed possibilities for offloading cognitive tasks to free up human bandwidth for tackling grand challenge problems.

Details of Moore's work were first reported by Oak Ridge National Laboratory through the DOE Science News Source.