US military dominates battlefield in a global AI revolution

Autonomous military machines are serving on active duty around the world, and the US is calling on public and private sector AI leaders to help keep the nation ahead of the field.

Last year, a research team from RAND conducted a survey that presented various scenarios on how the US military might use artificial intelligence. The team asked respondents to share their comfort level in using AI for these scenarios, which varied in terms of factors, such as the distance from the battlefield, the level of destructiveness, and the degree of human oversight over the AI algorithm.

The survey's results showed that most AI experts in the US do not oppose the fundamental mission of the US Department of Defense to use AI for various military applications.

“A global technology revolution is now underway,” says Antony Blinken, US Secretary of State, who also served as deputy assistant to then-President Barack Obama and national security advisor to Vice President Joe Biden from 2009 to 2013. Blinken returned to government as a foreign policy advisor for Biden's 2020 presidential campaign, having founded the consulting firm WestExec Advisors with fellow former Obama administration officials Michèle Flournoy, Sergio Aguirre, and Nitin Chadda in 2017.

“The world’s leading powers are racing to develop and deploy new technologies, like artificial intelligence and quantum computing, that could shape everything about our lives,” says Blinken. “From where we get energy to how we do our jobs, to how wars are fought. We want America to maintain our scientific and technological edge because it’s critical to us thriving in the 21st-century economy.”

The US federal government was estimated to have more than US$6bn in AI-related research-and-development projects in 2021 following the publication of President Donald Trump’s 2019 executive order, Maintaining American Leadership in Artificial Intelligence.

Human judgement still required for AI weapons

In January this year, the US Department of Defense (DoD) updated its directive on developing and deploying autonomous and semi-autonomous weapon systems, including those that incorporate artificial intelligence.

The core of the directive remains unchanged, but it was updated due to advances in technology, changes in the department's structure, and changes in the security environment. 

The directive requires a review by senior officials before developing and deploying any autonomous weapon systems that do not meet specific exemptions. DoD officials say it helps maintain “appropriate levels of human judgement” in their work.

DoD leaders believe the original 2012 directive remains fundamentally sound, but was due for some tweaks. "The updated directive is one part of a series of DoD policies that establish [good] governance surrounding military uses of autonomous systems and artificial intelligence," said a DoD official at the time of the update.

Hunting military obstacles to data sharing

The DoD is also leading an experiment that uses data, analytics, and artificial intelligence to inform solutions related to Joint All-Domain Command and Control (JADC2). The experiments – Global Information Dominance Experiments (GIDE) – are organised by the Chief Digital and Artificial Intelligence Office (CDAO) in partnership with the Joint Chiefs of Staff (JCS). They are the first to be held at this level and scale.

The experiment involves US military and civilian personnel from all service branches and multiple combatant commands. The aim is to identify barriers that inhibit data sharing across the Joint force, such as policy, security, connectivity, and user-interface issues.

The CDAO aims to improve access to data across the Joint force – from the strategic level to tactical warfighters – and show how this data, alongside analytics and AI, can improve Joint workflows in various missions, from globally integrated deterrence through to targeting and fires. The experiment's goal is to provide a rapid solution to the challenges faced by the Joint force and demonstrate how data can help achieve mission objectives more effectively.

“We want to stress-test our current systems and processes, introduce new technologies and approaches, and learn in an experimentation environment that replicates real-world operations,” says Col. Matthew Strohmeyer, GIDE V Mission Commander.

The CDAO will host four more iterations (GIDE V-VIII) of the experiment throughout 2023, aligned with the Joint Warfighting Concept and the JADC2 Implementation Strategy. By hosting these experiments, the DoD says it sets a precedent for joint, integrated experimentation enabled by data, analytics, and AI to advance their critical capabilities.

Lockheed Martin flies into AI history books

In the private sector, Lockheed Martin's VISTA X-62A recently made history as the first tactical aircraft flown by an artificial intelligence agent. VISTA, or Variable In-flight Simulation Test Aircraft, was developed by Lockheed Martin’s legendary Skunk Works and Calspan Corporation for the US Air Force Test Pilot School at Edwards Air Force Base in California.

VISTA is a modified F-16D Block 30 Peace Marble Il aircraft upgraded with Block 40 avionics and fitted with software that allows it to mimic the performance characteristics of other aircraft. The AI-powered VISTA platform can process vast amounts of data quickly and accurately, enabling it to optimise logistics, identify potential threats, and detect anomalies in data sets.

The recent 17-plus-hour flight by an AI agent on VISTA was made possible by a suite of advanced systems provided by Calspan and Lockheed Martin, including the VISTA Simulation System, Model Following Algorithm, and System for Autonomous Control of the Simulation. These systems enable VISTA to conduct the most advanced flight test experiments emphasising autonomy and AI.

"VISTA will allow us to parallelise the development and test of cutting-edge artificial intelligence techniques with new uncrewed vehicle designs," says Dr M. Christopher Cotting, US Air Force Test Pilot School director of research. "This approach, combined with focused testing on new vehicle systems as they are produced, will rapidly mature autonomy for uncrewed platforms and allow us to deliver tactically relevant capability to our warfighter."

Lockheed Martin is also developing live AI technologies in its Aegis Combat System, designed to deliver a superior solution and tactical advantage to the US Navy.

“The Aegis Combat System combines over 50 years of continuous evolution to provide rapid, modern updates to the warfighters,” says Joe DePietro, vice president and general manager of Naval Combat and Missile Defense Systems at Lockheed Martin. “Today, we are exploring how AI can provide faster system reaction time and decision support for operators and commanders.”