AI, Robots and the Future of Warfare: Discussions Among Mad Scientists
(Source: US Army; issued April 25, 2019)
AUSTIN, Texas --- The U.S. Army Training and Doctrine Command and the University of Texas at Austin hosted the Mad Scientist Conference at the university on April 24 and 25th. The Mad Scientist Conference brings together military, academia, and private industry experts in fields such as artificial intelligence (AI), robotics, ethics in future innovation, and the future of space.

This year's conference focused on disruption and the future operational environment. With the Army's effort to modernize the force, it is critical for collaboration between the Army and the brightest minds of technological innovation.

"Mad Scientist and Army Future Command are two sides of the same modernization coin," said Lt. Gen. James Richardson, deputy commanding general of Army Futures Command. "We need to tap into America's unique culture of innovation. That's why we're here in Austin. AFC is an opportunity for collaboration with the best minds in the world in academia and industry."

Collaboration today to solve the complex problems of tomorrow's battlefields requires significant imagination to predict possibilities.

"The future of warfare will be both familiar and utterly alien," Richardson said.

With the development of evolving artificial intelligence and robotics, Mad Scientists discussed the applications they have on future warfare.

"When technology is proliferated down to the battlefield, what happens?" asked Robert Work, senior counselor for defense and distinguished senior fellow for defense and national security at the Center for a New American Security. "We'll inevitably go to more unmanned systems."

While wars today feature manned combat vehicles, the Mad Scientists suggest wars of the future may be fought by drones and AI-controlled machines. Work referenced the Army's next generation combat vehicle currently in development that has the potential to be optionally manned.

One way future vehicles can operate without a human crew is using AI.

"How do we make autonomous systems behave in a trustworthy fashion?" asked Dr. Maruth Akella, professor of aerospace engineering and engineering mechanics at UT-Austin.

A primary goal of AI and robotics is full autonomy to perform increasingly complex tasks. The Mad Scientists questioned how to establish ethics and human oversight for automated machines used on complex battlefields where non-combatants, enemy forces and partner forces are intermingled in real-time, dynamic domains.

The discussions examined how much autonomy should autonomous machines have in military operations.

"How much human control do we want or need to have over these autonomous systems?" asked Dr. Paul Zablocky, program manager for the strategic technology office of the Defense Advanced Research Projects Agency.

To further understand the implications of autonomous machines in the operational environment, the conference speakers discussed how AI learns and how humans are involved in the AI-learning process.

"We need to look at integrated human-in-the-loop systems," said Dr. Garrett Warnell, a research scientist with Army Research Lab. "When robots are becoming autonomous, they need a lot of human interaction. They slowly depend less and less on humans and become more autonomous."

If robotics are considered for warfare in the future, Work said we must pursue systems with tele-operated capabilities. Additionally, the panelists strongly emphasized that robotics must be disposable, which opened the conversation to how much these technologies might cost. Work pointed out that China could pass the US in absolute GDP in about 10 years.

"The U.S. cannot spend our way back to military dominance," said Work. "That means that we have to out-think, out-innovate, and out-maneuver our competitors."

The opportunity to collaborate, out-think and out-innovate is the reason that Army Futures Command was created and based in Austin amongst a variety of tech companies, start- ups, and innovators.

-ends-







prev next