Page 17 - ADT JANUARY - FEBRUARY 2023 Online Magazine

U.S. - U.K. JOINT MISSION AI training and retraining on deployed tacti-
For the first time, the AFRL, in partnership with the U.K.’s Defence cal high-performance computers to deliver
Science and Technology Laboratory, or Dstl, recently demonstrated mission-specific AI. It also included a test and
state-of-the-art artificial intelligence technology at two major evaluation process to ensure that the AI devel-
back-to-back military exercises. oped was robust and trusted. “The whole team
across both nations came together to deliver
A team of 30 AI and autonomy experts from the U.K. and U.S. this landmark trial for the collaboration,” said
deployed as a joint taskforce to the Project Convergence 22, or Todd Robinson, U.K. AI toolbox lead. “By deploy-
PC22, experiment at the U.S. National Training Centre at Fort Irwin, ing our AI taskforce to PC22, we learned what
California. Later, a subset of the taskforce reconvened at the British this technology would mean to the warfighter
Army’s Salisbury Plain Training Area in Wiltshire, England, taking and identified further challenges, which require
lessons learned from PC22 and rapidly applying AI into a new oper- research and development to enhance a future
ational environment as part of the Dstl HYDRA project’s Integrated operational capability. It is important we deploy
Concept Evaluation, or ICE. AI into trials more regularly to drive the matura-
tion and operationalization of AI.”
Both exercises addressed the challenge of making AI and auton-
omy agile, adaptable, trustworthy and accessible to warfighters, To make AI understandable and enable users
albeit under different U.S. and U.K. military use cases. The goal was to select appropriate algorithms for missions,
to deliver mission specific AI that can be deployed to meet the the team developed model cards with clear
ever-changing mission conditions and needs of warfighters. descriptions of the algorithms’ strengths and
weakness. “It is becoming more and more crit-
Led by the U.S. Army Futures Command, PC22 is the Joint Force ical to be able to adapt AI to meet changing
experimenting with speed, range, and decision dominance to mission requirements, operating environments
achieve overmatch and inform the Joint Warfighting Concept and and accelerated decision timelines in-mission,
Joint All Domain Command and Control. A campaign of learning, it all while ensuring it is trusted and understand-
helped the forces leverage a series of joint, multi-domain engage- able to the military users,” said Dr. Lee Seversky,
ments to integrate artificial intelligence, robotics and autonomy AFRL project lead. “The joint AI toolbox, with
to improve battlefield situational awareness, connect sensors with its ability to adapt and deliver AI for different
shooters, and accelerate the decision-making timeline. joint military missions is critical. AI flexibility and
During the PC22, the joint AI taskforce deployed a U.K.-U.S. AI tool- speed is key to moving us towards this goal.”
box for the first time, which enabled warfighters across the coalition
to select the best AI tools for the missions, said Dr. Lee Seversky, AFRL According to AFRL and Dstl, autonomy and AI
project lead and the U.S. AI toolbox lead. are critical to future warfighters.

The toolbox used data collected from U.K.-U.S. uncrewed ground “These are rapidly emerging technologies that
vehicles and uncrewed aerial vehicles, or UAVs, along with rapid we must be able to understand and grasp to
ensure that our warfighters have the tools they
need to win on the battlefields of the future,”
said John Godsell, Dstl U.K. autonomy pro-
gramme manager. “It was hugely exciting to
be part of Project Convergence 2022, an exper-
iment at an epic scale.”

In December 2022, the AI toolbox was deployed
on U.K. platforms as part of the Dstl HYDRA
project’s ICE trials on Salisbury Plain. ICE4
demonstrated that U.K.-U.S. developed algo-
rithms from the AI toolbox could be deployed
onto a swarm of U.K. UAVs and retrained by the
joint AI taskforce at the ground station and the
model updated in flight, a first for the U.K. This
demonstrated how the AI toolbox adapts to new
data sources, platforms and operating locations
to provide rapid updates to the AI deployed
onto autonomous systems.

“ICE4 has enabled us to consider the prac-
ticalities of how AI could be used to support
swarming UAS operations in contested envi-
ASIAN DEFENCE TECHNOLOGY JANUARY-FEBRUARY 2023 | 17