(This Article was updated on February 6, 2023, to include statements from DARPA officials.)
The U.S. Department of Defense has quietly launched a new program to develop the ability to unleash thousands of autonomous land, sea, and air drones capable of overwhelming and dominating an enemy’s area defenses.
Details of the project are largely shrouded in secrecy. However, language in recently updated pre-solicitation documents suggests the development of massive autonomous vehicle swarms likely has a specific focus on deterring or defeating a potential Chinese invasion of Taiwan.
According to documents, the effort will be managed by the Strategic Technology Office of the Defense Advanced Research Projects Agency (DARPA).
And while DARPA has yet to formally announce the new program, DARPA confirmed to The Debrief that the project will be operating under the moniker “Autonomous Multi-Domain Adaptive Swarms-of-Swarms” program, or “AMASS.”
“DARPA’s singular and enduring mission is to make pivotal investments in breakthrough technologies for national security,” said a DARPA spokesperson via email.
“The DARPA AMASS program is exploring the use of swarms-of-swarms to conduct military operations in highly contested environments that pose great risk to our Soldiers, Sailors, Airmen, and Marines. It is meant to inform future military programs of record and not to be a military program of record itself.”
AMASS is hardly the first, or only, DARPA project currently underway leveraging autonomous drone swarm technology to give the U.S. military an advantage on future battlefields.
For the past six years, the OFFensive Swarm-Enabled Tactics Program (OFFSET) has been perfecting the ability for up to 250 autonomous aerial and land-based drones to collectively descend upon urban environments to assist Army ground forces.
Typically, DARPA projects, like OFFSET, involve testing the feasibility of various high-risk, high-reward “breakthrough” technologies. The fruits of these labors are then farmed out to interested military branches for further development or incorporated into existing systems.
However, AMASS seems unique in that DARPA already appears to have a specific mission in mind for using “swarms-of-swarms” of autonomous drones.
“Today, our peer-state adversaries could invade their neighbors with little warning given their time-distance-mass advantage,” reads a Broad Agency Announcement (BAA) updated by DARPA on January 30.
“Adversary A2/AD [anti-access/area denial] bubbles with sophisticated air defense, indirect fires, precision weapons, and Intelligence, Surveillance, and Reconnaissance (ISR) capabilities would severely limit Joint Service and Coalition operations under the bubble.”
DARPA hints at AMASS having a definite mission by noting, “this program will be oriented around demonstrating this capability for some specific regional Theater-level scenarios.” The BAA further reiterates, “This program will be experimentation and scenario-focused with a specific regional emphasis.”
The solicitation does not name any particular “peer-state adversaries.” However, mentions of invading neighboring states with little warning and AMASS having a “specific regional emphasis” are highly suggestive of Pentagon concerns over a potential Chinese invasion of Taiwan
Less ambiguous is AMASS’s ultimate goal.
“Central to the AMASS program is the ability to plan and execute missions that utilize thousands of autonomous entities in the degradation or defeat of adversary A2/AD capabilities.”
Documents show that the goal of AMASS is to develop the capability to launch and command thousands, perhaps tens of thousands, of autonomous drones, working together to destroy a multitude of an enemy’s defenses, including air defense, indirect and precision weapons, and intelligence, surveillance, and reconnaissance platforms.
“The current vision is that low-cost swarms with diverse sensors and kinetic and non-kinetic effectors would primarily be pre-positioned forward and launched remotely, providing rapid response and adaptability to overcome the adversary’s time-distance-mass advantage.”
The AMASS program is expected to include swarms of multi-domain autonomous vehicles, including Unmanned Aerial Vehicles (UAVs), Unmanned Surface Vehicles (USVs), and Unmanned Ground Vehicles (UGVs).
In a January 10 Q&A session with potential industry partners, a DARPA official indicated that Unmanned Underwater Vehicles (UUVs) would also be considered for the AMASS program.
“We are not developing new autonomous platforms under the program; we are leveraging existing, small, low-cost platforms,” explained a DARPA spokesperson.
“AMASS is focused on developing command and control systems that plan and execute missions using thousands of autonomous entities to degrade or defeat an adversary’s anti-access/area denial (A2/AD) capabilities and will explore how these swarms of autonomous vehicles would interact with one another.”
Solicitation documents make it clear that AMASS is not some curiosity-driven research program. Instead, the Pentagon seems to believe it needs a large-scale robotic army sooner rather than later.
The award winner of the AMASS contract is expressly prohibited from engaging in research that “primarily results in evolutionary improvements to the existing state of practice.” Instead, a contractor should focus on “innovative approaches that enable revolutionary advances in science, devices, or systems.”
DARPA notes that the successful completion of the AMASS contract will likely require an industry partner to leverage a wide variety of technologies and incorporate “lessons learned” from the agency’s System-of-Systems Enhanced Small Unit (SESU) program.
Initial funding for AMASS is anticipated to be $78 million, with the award expected to go to a single private contractor. DARPA, however, cautions that “based on the complexity of this program, proposers are highly encouraged to seek teaming arrangements with subcontractors who have relevant products and expertise.”
Following federal law, the AMASS solicitation is publicly available through the U.S. government’s online e-procurement system, SAM.gov. However, much of the program is already cloaked in secrecy.
“The effort being solicited by this BAA is classified SECRET or otherwise involves access to or generation of classified information,” the AMASS solicitation notice reads. According to records, the DoD issued a formal security classification guide for AMASS before publishing a pre-solicitation notice on November 22,2022.
“The goal is for hardware, software, and raw data collected to be unclassified to the maximum extent possible,” explained a DARPA official during a Q&A session for prospective vendors. However, test scenarios, experimental results, and live experiments are expected to be conducted at the classified level.
According to DARPA, at least some live testing scenarios for AMASS will occur outside the contiguous United States.
The inherent secrecy of AMASS’s development will likely exacerbate existing concerns surrounding emerging technologies like artificial intelligence and autonomous drones for military use. Particularly the unease of robotic systems having the independent ability to decide when to pull a trigger or drop a bomb.
In a 2017 letter to the United Nations, 116 AI experts, including Tesla and SpaceX founder Elon Musk, expressed concerns about the dangers of developing military hardware capable of “thinking” for itself and lacking any need for human guidance.
“Once developed, they will permit armed conflict to be fought at a scale greater than ever and at timescales faster than humans can comprehend,” the letter said.
Current Department of Defense policy states autonomous or semi-autonomous weapons systems must be “designed to allow commanders and operators to exercise appropriate levels of human judgment over the use of force.”
Citing a 2018 U.S. government white paper, the Congressional Research Service found that vagueness in DoD policy has opened the door for military commanders to authorize autonomous drones to use force before deployment, thus still giving the machines the ability to choose when to pull the trigger.
According to DARPA, the intent of the AMASS program isn’t to develop autonomous drones with capability to independently execute lethal missions.
“The central technical challenge of this program is to design human-in-the-loop planning and establish criteria to bound the autonomous operations. This includes establishment of geofences for allowed operations, required confidence levels and permissions before taking action, and automated mission termination,” a DARPA spokesperson explained.
“Key to this will be developing software that fails into safe modes regardless of communications status and the ability of humans to intercede and set limits. The formal development and fielding of capabilities that are informed by DARPA’s AMASS program will comply with Department of Defense (DoD) Directive 3000.09, “Autonomy in Weapons Systems” dated January 25, 2023.”
Ultimately, despite any concerns, the use of militarized autonomous technologies is likely no longer a question of “if” but a certainty of “when.”
Speaking with The Atlantic, Samuel Bendett, an expert in Russian weapons at the Center for Naval Analyses, characterized drone swarms as “the holy grail” of defense technologies.
“This is what everybody’s working towards,” Bendett said. “By everybody, I mean advanced countries and advanced militaries hoping to utilize swarm technologies.”
“So the list is short, but it’s slowly growing.”
Tim McMillan is a retired law enforcement executive, investigative reporter and co-founder of The Debrief. His writing typically focuses on defense, national security, the Intelligence Community and topics related to psychology. You can follow Tim on Twitter: @LtTimMcMillan. Tim can be reached by email: firstname.lastname@example.org or through encrypted email: LtTimMcMillan@protonmail.com