AI robots could assist soldiers in future: Study

Scientists in the US are developing artificial intelligence (AI) systems that could help robots assist soldiers in the battlefield in future.

For the research, published in the journal Science Advances, team looked at soldier brain activity during specific tasks for ways to incorporate AI teaming to dynamically complete tasks.

According to Jean Vettel, a senior neuroscientist at the Army Research Laboratory (ARL) in the US, technologies that can predict states and behaviours of the individual soldier may help create a more optimised team.

The work between ARL and the University at Buffalo is looking at ways the dynamics and architecture of the human brain may be coordinated to predict such behaviors and consequently optimize team performance.

"In military operations, soldiers perform multiple tasks at once. They're analysing information from multiple sources, navigating environments while simultaneously assessing threats, sharing situational awareness, and communicating with a distributed team," said Vettel.

"This requires soldiers to constantly switch among these tasks, which means that the brain is also rapidly shifting among the different brain regions needed for these different tasks," he said.

"If we can use brain data in the moment to indicate what task they're doing, AI could dynamically respond and adapt to assist the Soldier in completing the task," he added.

To achieve this future capability, the researchers first sought to understand how the brain coordinates its different regions while executing a particular task. They used a computational approach to understand how this may be characterised to inform the behavioral prediction.

To complete the study, researchers mapped how different regions of the brain were connected to one another in 30 different people via tracts of tissue called white matter.

Scientists converted these maps into computational models of each subject's brain, and used computers to simulate what would happen when a single region of a person's brain was stimulated.

They then used a mathematical framework, which they developed, to measure how brain activity became synchronised across various cognitive systems in the simulations.

"While the work has been deployed on individual brains of a finite brain structure, it would be very interesting to see if coordination of Soldiers and autonomous systems may also be described with this method, too," said Javier Garcia, an ARL neuroscientist.

"Much how the brain coordinates regions that carry out specific functions, you can think of how this method may describe coordinated teams of individuals and autonomous systems of varied skills work together to complete a mission," Garcia said.