The concern, critics say, is not limited to weapons that can select and engage targets on their own.

At the hearing, titled “Artificial Intelligence in the Military Domain: Implications for Human Rights,” witnesses told members of the Tom Lantos Human Rights Commission that AI can influence human decisions leading to a strike while leaving the individual formally responsible for the final call.

“The approval button alone does not demonstrate the control,” Anna Mysyshyn, a Ukrainian AI-governance researcher, told lawmakers.

The issue comes as the Pentagon pushes to use AI to accelerate military decision-making. The Pentagon’s 2023 AI adoption strategy identified “fast, precise and resilient kill chains” as a desired outcome, while its 2026 AI strategy calls for becoming an “AI-first” warfighting force by redesigning military workflows and operational concepts around current AI capabilities.

Mysyshyn questioned whether a person supervising multiple AI-enabled systems can recognize an incorrect recommendation and intervene fast enough before harm occurs. She cited Ukraine as an example of AI compressing decisions from hours to minutes.

In many scenarios, a human making the final decision does not necessarily have enough information, time or the ability to challenge the system’s recommendation, she said.

The Defense Department’s 2023 directive on autonomy in weapon systems requires autonomous and semi-autonomous weapons to allow commanders and operators to exercise appropriate levels of human judgment over the use of force.

Wednesday’s testimony focused on a broader question: whether that judgment remains meaningful if AI is influencing targeting faster than a human can authenticate.

Amanda Klasing, national director of government relations and advocacy at Amnesty International USA, warned that AI can compress the military “kill chain” in ways that make it difficult to determine whether meaningful human discernment actually occurred before a strike.

Joseph Chapa, a former military officer and scholar of military ethics, argued that the debate should not focus solely on how autonomous a system is. He urged policymakers to consider whether an AI system relies on statistical inference, what happens when it fails in a particular operational environment and whether its decisions can be audited afterward.

Statistical AI systems can fail differently from traditional software, Chapa said, including producing substantially different outputs from relatively small changes in inputs. More complex systems can also make it difficult to reconstruct how a particular recommendation was produced after the fact.

He called for greater AI education across the military, including specialized training for personnel working with targeting and intelligence systems.

Witnesses on Wednesday also pointed to a larger accountability problem: determining what happens when an AI-assisted decision results in civilian harm.

The Defense Department established its Civilian Harm Mitigation and Response Action Plan in 2022, including a Civilian Protection Center of Excellence intended to serve as a hub for analysis and training. Federal law requires the Pentagon to operate the center.

But a Defense Department inspector general review released in May assessed the department’s implementation of the civilian-harm framework and found problems including the loss or reassignment of personnel, the end of funding for a civilian-harm data-management platform and the failure of a senior steering committee to continue meeting.

That matters in the case of AI because, as Klasing noted, the military currently has limited ability to determine whether AI may contribute to improper targeting of civilians.

On the first day of the U.S.-Iran war, a Tomahawk cruise missile struck Shajareh Tayyebeh elementary school in Minab, southern Iran. At least 168 people were killed — more than 100 of them under the age of 12, according to UN and Iranian officials.

The school building sat fewer than 100 yards from a long-time Islamic Revolutionary Guard Corps naval installation and was previously located within the IRGC compound perimeter until a wall appeared between 2013 and 2016, according to an analysis of satellite imagery by Amnesty International.

By the time the U.S. and Israel launched their first strikes on Feb. 28, the school had been established several years prior. It was active on social media and had its own website, a Reuters investigation found.

More than 120 House Democrats sent a letter to the Pentagon on March 12, demanding clarity on whether artificial intelligence, “including the use of the Maven Smart System, [was] used to identify the Shajareh Tayyebeh school as a target.”

U.S. military investigators in the wake of the incident stated that the strike was likely carried out by U.S. forces. The Pentagon, however, has still not released its investigative findings.

Klasing proposed adding information about AI involvement to the Pentagon’s annual civilian-casualty reports required under Section 1057 of the National Defense Authorization Act. The existing reporting requirement calls for information about civilian casualty mitigation efforts and investigations, but does not specifically require the department to identify whether AI contributed to an incident.

Steve Feldstein, a senior fellow at the Carnegie Endowment for International Peace, said the Pentagon should collect data from military operations and conduct after-action reviews to determine whether AI-assisted systems accurately identified intended targets and whether errors contributed to civilian casualties.

“A bigger question is to what extent is the Pentagon collecting data and verifying the target accuracy of all these 1000s of strikes?” Feldstein said.

Feldstein also cautioned against treating AI itself as the source of every potential problem. AI can be used to help verify intelligence, detect changes in imagery and prioritize defensive responses to threats such as drones and ballistic missiles.

But if a human-influenced system is built around assumptions that certain demographic traits make someone a lawful target, Mysyshyn said, AI would not correct that underlying premise.

Feldstein similarly pointed to the rules humans establish before an AI system is used. If commanders determine that a particular level of civilian harm is acceptable in pursuit of a military objective, he said, the machine is operating within a human-defined framework.

The challenge is particularly acute when systems operate faster than people can meaningfully assess recommendations.

The commission opened Wednesday’s hearing with an example from Iraq. In 2004, Army squad leader Nick Ellison encountered two young men firing a machine gun from a car during a census patrol in Kirkuk.

Ellison made eye contact with one of the men, saw panic and ordered his soldiers to hold fire. The men turned out to be brothers of a newly married groom, and the gunfire was celebratory.

Mysyshyn said testing should examine bias, reliability, civilian protection, realistic time pressure, whether operators can interrupt systems and whether decisions can be traced throughout an AI system’s lifecycle.

Chapa said some AI systems may be difficult to audit after the fact, while Deborah Brown of Human Rights Watch said information about military AI use has often emerged through investigative reporting and whistleblowers rather than government disclosures.

The growing challenge, according to witnesses, is no longer making a machine pull a trigger, but whether the human who approves the trigger pull has enough time, information and authority to make a meaningful choice before disaster.

Military Times reporter Katie Livingstone contributed to this report.

The department premiered WarMatrix for its inaugural use at the March 27 GE 26 Benchmark Wargame, marking the system’s move from development into operational capacity, according to a Tuesday Air Force release.

WarMatrix, described by the force as an “active wargaming environment,” is an AI-powered system that integrates existing models, data and workflows while expediting analysis.

The Air Force at the end of 2025 said it was looking for technology capable of producing simulations 10,000 times faster than real time.

WarMatrix is a “human-machine teaming system” meant to keep human judgment integral to planning and decision making, according to the release.

The use of WarMatrix during the event served as the system’s initial operating concept evaluation, signaling a change in how the Air Force conducts operational analysis and wargaming.

“Designed by wargamers for wargamers, WarMatrix provides transparency, auditability and speed, enabling decision-makers to better understand assumptions, outcomes and tradeoffs,” the statement reads.

Air Force leaders portray WarMatrix as an evolution in wargaming rather than a replacement, and the release says that the use of WarMatrix provided a more “connected and traceable wargaming process.”

It also said that the system’s design allowed for faster scenario development, repeat findings and increased collaboration with joint and coalition partners.

The two-weekslong event, hosted at Systems Planning and Analysis in Alexandria, Virginia, was attended by more than 150 people, including technical experts, Air Force leadership and allied planners.

Attendees during the event fulfilled more than six 24-hour “game-time moves” that balanced physics- and simulation-based models to ensure outcomes were realistic.

The defense technology firm introduced the system, known as MAYHEM 10, which expands upon its Switchblade family.

The Army in February announced a $186 million purchase that includes two variants of Switchblade one-way attack, or “kamikaze,” drones: the Switchblade 600 Block 2 variant and the Switchblade 300 Block 20 variant.

The difference is that MAYHEM 10 is multifunctional, meaning it can perform tasks in addition to striking. The new system can carry a 10-pound payload and has a range of over 62 miles, per the release.

The system is capable of 50 minutes of endurance, with a launch assembly that can be done in under five minutes, the statement says. It can also be launched from the air, ground or maritime platforms.

“By integrating advanced autonomy, multi-domain payloads, and rapid adaptability, we empower our forces to sense, disrupt, and strike with precision — even in the most contested environments,” Wahid Nawabi, AeroVironment’s chairman, president and chief executive officer, said in the statement.

Last year, U.S. soldiers tested the Switchblade 600 system, which has a range of 27 miles and is designed to engage a target using onboard cameras.

The Switchblade 300 Block 20, unlike the heavier 600 variant, is small enough to be carried in a backpack. For the first time, according to a February AeroVironment announcement, it will come equipped with an Explosively Formed Penetrator, a deadly warhead that is made to penetrate armored vehicles.

The U.S. military is paving the way for the regular deployment of high-energy laser weapons on American soil for air defense amid the expanding threat of low-cost weaponized drones.

The Federal Aviation Administration and the U.S. Defense Department have reached a “landmark safety agreement” regarding the use of laser weapons to counter unauthorized drones at the US-Mexico border following a safety assessment that concluded such countermeasures “do not pose undue risk to passenger aircraft,” the FAA announced Friday.

The assessment and resulting agreement were the direct result of two laser incidents along the southern border of Texas in February, which prompted the FAA to abruptly close nearby airspace amid concerns over the potential impact on civilian air traffic. The incidents involved the U.S. Army’s 20 kilowatt Army Multi-Purpose High Energy Laser (AMP-HEL), a vehicle-mounted version of defense contractor AV’s LOCUST Laser Weapon System.

In the first incident, U.S. Customs and Border Patrol personnel used an AMP-HEL on loan from the Pentagon to engage an unidentified target near Fort Bliss, triggering an airspace shutdown above El Paso on Feb. 11. In the second, U.S. military personnel used an AMP-HEL near Fort Hancock to neutralize a “seemingly threatening” drone that turned out to belong to CBP, spurring another shutdown on Feb. 27.

“Following a thorough, data-informed Safety Risk Assessment, we determined that these systems do not present an increased risk to the flying public,” FAA administrator Bryan Bedford said in a statement. “We will continue working with our interagency partners to ensure the National Airspace System remains safe while addressing emerging drone threats.”

The “first of its kind” safety assessment, conducted in early March by the FAA and the Pentagon’s Joint Interagency Task Force 401 (JIATF-401) counter-drone organization at White Sands Missile Range in New Mexico, reportedly yielded two significant conclusions: 1) the LOCUST’s automatic shutoff mechanism will consistently prohibit the system from firing under unsafe circumstances, a point that AV executives have emphasized in recent weeks, and 2) in the event of a system failure, the laser beam itself cannot inflict catastrophic damage even on aircraft flying at its maximum effective range, let alone those at cruising altitudes.

Here’s how Aaron Westman, AV senior director for business development, described the LOCUST’s safety protocols in a company blog post on March 23:

Every time an operator presses the “fire” button, the system runs through a series of automated checks. Some examples include:

• Is the laser pointing away from protected “keep-out” zones?
• Are all internal subsystems operating within safe parameters?
• Is the system properly locked onto a target?
• Are safety interlock switches engaged?
• Are all software safety checks satisfied?

Each of these checks acts as a safety “vote.”

If any subsystem registers a “no vote,” the laser simply will not fire. An operator can press the trigger — and nothing happens. The system refuses to engage until all conditions are verified as safe.

These automated safeguards are built into both the hardware and the software of the system.

Here’s how DefenseScoop described the LOCUST’s potential effects on passing airframes based on an account from Army Col. Scott McLellan, JIATF-401 deputy director, of the testing at White Sands:

McLellan said the evaluation involved “localized” firing of the AMP-HEL from various distances at the fuselage of a Boeing 767 airliner that testers lugged on to White Sands to assess the system’s damaging effects, “or lack thereof” on aircraft material. He said it aimed to “disprove some myths” about the capability, noting “that energy clearly dissipates over time and space and doesn’t have the effect everyone thinks it does as far as lasers are concerned.”

A JIATF 401 spokesperson said the laser was fired at its “maximum effective range for up to 8 seconds” at the grounded fuselage, “demonstrating that even at full intensity, the laser caused no structural damage to the aircraft.”

As drone warfare spreads beyond distant conflicts, laser weapons are an increasingly attractive domestic countermeasure. While kinetic interceptors and electronic warfare may be considered suitable for chaotic battlefields, their potential for collateral effects makes them far too risky for consistent domestic applications. And even if collateral damage wasn’t a concern, expending expensive missiles on the 1,000 cartel-operated drones that cross the border with Mexico monthly is economically unsustainable, especially for a Pentagon that’s already rapidly burning through munitions as part of Operation Epic Fury against Iran. On paper, the argument seems obvious: Why not save those critical interceptors for high-end threats overseas and let domestic laser emplacements, with their deep magazines and minimal cost-per-shot, pull counter-drone duty at home?

Using laser weapons for domestic air defense wouldn’t be unprecedented. France deployed two 2 kw High Energy Laser for Multiple Applications – Power (HELMA-P) systems to secure the airspace over the country’s Île-de-France region during the 2024 Paris Olympics and Paralympics. This past September, China’s People’s Liberation Army deployed several laser weapons across Beijing during a major military parade marking the 80th anniversary of Japan’s defeat at the end of World War II. As of January, the U.K. Ministry of Defense was reportedly drawing up plans to build a domestic laser screen, albeit composed of lower-power laser dazzlers, to protect military installations and other critical infrastructure. The Pentagon has even already considered laser weapons to reinforce the airspace above Secretary of Defense Pete Hegseth and Secretary of State Marco Rubio’s residences at Fort McNair in Washington following a series of unauthorized drone incursions there.

Indeed, there’s a distinct possibility that laser weapons could see increasing domestic applications amid the U.S. military’s growing appetite for novel drone defenses. On April 2, JIATF-401 announced that it had funneled $20 million in counter-drone systems like the Dronebuster EW handset and Smart Shooter computerized riflescope to the U.S.-Mexico border in just four months.

Days later, the task force announced $100 million to enhance counter-drone capabilities for the 2026 FIFA World Cup starting in June “to protect stadiums and fan zones in 11 cities across nine states,” part of larger $600 million surge in counter-drone systems that also allocated $158 million to “defend the nation’s highest-priority defense critical infrastructure.”

With the Pentagon asking for $580 million in R&D funding just for JIATF-401 in its fiscal year 2027 budget request (and potentially $800 million in procurement cash), the task force appears poised to explore any and all possible solutions to the drone problem — and operationally, the FAA-Pentagon safety agreement helps establish laser weapons as a viable option.

That said, the safety agreement on its own is unlikely to open the floodgates for a sudden spate of laser weapon deployments along the U.S.-Mexico border, let alone for major events like the World Cup or critical infrastructure just yet. First, the agreement doesn’t appear to clarify who has final say in authorizing a laser engagement when U.S. military, CBP and FAA jurisdictions overlap — the precise ambiguity that yielded February’s airspace closures and, until resolved, will complicate future engagements during a fast-moving crisis. Second, the U.S. military’s arsenal of operational laser weapons is currently limited despite a stated goal of rapidly fielding new systems at scale within three years. Even with clear plans to surge directed energy research and development for homeland defense under President Donald Trump’s “Golden Dome for America” missile shield, the age of sleek beam directors quietly standing watch along the US-Mexico border remains a long way off.

The FAA agreement may end up laying the foundation for a true domestic laser air defense architecture — a “Laser Dome” in all but name. Whether the U.S. military actually builds it, however, will depend not just the Pentagon’s promise to deploy laser weapons at scale, but whether Washington can finally sort out who’s in charge when a beam crosses into civilian airspace.

The Pentagon announced the CAMP program last April with the release of a solicitation calling for a new class of XL-AUV, or extra large underwater vehicle.

The solicitation called for a large underwater vessel that could maneuver in GPS-denied environments, dive over roughly 656 feet (200 meters) deep and drop “various payloads to the sea floor.” It specified that the vessel had to stay underwater for prolonged periods and be able to be transported and recovered easily by commercially available freight and logistics equipment.

Anduril was chosen for the CAMP program after wrapping the longest-ever XL-AUV demonstration, per a company statement. It will conduct an extended demonstration of its Dive-XL platform within four months of contract award.

“The subsea domain currently is patrolled and serviced by a very small number of exquisite capabilities,” Dr. Shane Arnott, senior vice president of Anduril’s maritime division, told reporters at a March 11 roundtable. “The Pacific and also the High North, which is the new fight that’s starting around the Arctic, are water-based fights. So it’s a no-brainer that robots are needed to supplement the crewed systems.”

The Dive-XL meets many of the desired features noted in the original CAMP solicitation. The vehicle is propelled by an all-electric powertrain that allows it to speed through the depths without needing to break the surface and can travel in excess of 2,000 nautical miles.

It has a highly flexible design that be modified for different mission types, including reconnaissance, and can carry up to three payload modules at once. These loads can include smaller unmanned underwater vehicles, effectively making the Dive-XL an type of mothership for even smaller drones that can conduct surveillance or strike underwater.

Additionally, the Dive-XL platform fits into commercial freight containers and can be transported via train or trucks, which corresponds to the Navy’s demand for a modular, easy-to-use system.

Anduril’s selection for the project follows its formal adoption last September by the Royal Australian Navy to deliver a fleet of AI unmanned submarine called Ghost Sharks within the next three years.

Video released Wednesday shows the next-gen munitions being launched in open desert terrain from M142 High Mobility Artillery Rocket Systems as part of Operation Epic Fury. The operation against Iran has seen a variety of precision munitions launched from land, sea and air, according to CENTCOM.

“I just could not be prouder of our men and women in uniform leveraging innovation to create dilemmas for the enemy,” Adm. Brad Cooper, CENTCOM commander, said in a release.

The PrSM is enhanced with GPS capabilities to navigate in flight and can accurately strike targets from around 250 miles away. Its warhead is designed to deliver a fragmentation effect when it explodes, increasing power on impact. The missile, meanwhile, is reportedly built to withstand turbulent in-flight conditions.

The PrSM is among a wide array of advanced technological assets the U.S. military has deployed during its ongoing operations in Iran.

Other notable weapon deployments in the ongoing conflict include Patriot Interceptor Missile Systems and THAAD Anti-Ballistic Missile Systems. LUCAS one-way attack drones were also used for the first time in combat by the U.S. Special Operations Command-led Task Force Scorpion Strike.

Specific details about the targets engaged by the PrSMs have not been provided as of publication.

Prior to its combat debut, the U.S. Army was working with Lockheed Martin to ramp up production capacity for the munition, Defense News reported last October.

The Army tested its short-range and long-range capabilities on White Sands Missile Range in April and September of last year. Tests saw the munitions fired from HIMARS systems as well as the M270A2 Multiple Launch Rocket System, or MLRS.

Operation Epic Fury has seen over 2,000 military targets across Iran destroyed, the Pentagon said, including a naval frigate that was sunk in the first torpedo kill by a U.S. Navy submarine since World War II.

The Defense Department’s Joint Interagency Task Force 401 announced the agreement with Perennial Autonomy on February 4, which will see deliveries of the Bumblebee V2 start next month.

The Bumblebee V2, a first-person view drone with four rotors, acts as its own drone hunter-killer system. Advanced artificial intelligence software enables it to identify other unmanned aerial vehicles and engage them in combat. It attacks other drones and knocks them out of the sky with a direct physical collision.

“This decision puts kinetic [counter-small] UAS capability into the hands of our troops immediately,” Army Brig. Gen. Matthew Ross, JIATF-401 director, said in a statement.

“The Bumblebee V2 provides a cost-effective, reliable interceptor that can neutralize threats without endangering our own forces or surrounding infrastructure. On the modern battlefield, where drones are a constant threat, having a low-collateral kinetic option is not just an advantage; it is increasingly becoming essential for protecting our forces.”

Direct physical strikes on unmanned aerial vehicles, known as “hard kills,” can produce a blast radius and scatter sharp debris when munitions are used.

With a drone-on-drone strike, the Bumblebee minimizes these risks. Having one flying drone tackle another eliminates the need for bullets and potentially harmful explosions. This potential could see the Bumblebee used efficiently in urban areas.

Additionally, the system is compliant with updated regulations in the National Defense Authorization Act, which has strict requirements for the use of foreign components. Drones with NDAA compliance, such as the HANX drone developed by the U.S. Marines, can be adopted for military use with immediate effect.

The Army’s Global Response Force will assess the system in support of the Lt. Gen. James M. Gavin Joint Innovation Outpost, a new facility opened at Fort Bragg on Jan. 3 to bring together soldiers, academics and private industry to develop and test new technology.

The vulnerability to such “prompt injection attacks” exists because large language models, the backbone of chatbots that digest hordes of user text to generate responses, cannot distinguish between malicious and trusted user instructions.

“The AI is not smart enough to understand that it has an injection inside, so it carries out something it’s not supposed to do,” Liav Caspi, a former member of the Israel Defense Forces cyberwarfare unit, told Defense News.

In effect, “an enemy has been able to turn somebody from the inside to do what they want,” such as deleting records or biasing decisions, according to Caspi, who co-founded Legit Security, which recently spotted one such security hole in Microsoft’s Copilot chatbot.

“It’s like having a spy in your ranks,” he said.

Former military officials say that, with greater reliance on chatbots and hackers backed by China, Russia and other nations already instructing Google’s Gemini, OpenAI’s ChatGPT and Copilot to create malware and fake personas, a prompt injection that orders the bots themselves to copy files or spread lies looms near.

Microsoft’s annual digital defense report, released last month, for the first time said, “AI systems themselves have become high-value targets, with adversaries amping up use of methods like prompt injection.”

What’s more, the problem of prompt injection has no easy solution, OpenAI and security researchers say.

An attack simply involves hiding malicious instructions — sometimes in white or tiny text — in a chatbot or content that the chatbot reads, such as a blog post or PDF.

For example, a security researcher demonstrated a prompt injection attack against OpenAI’s new AI-based browser, ChatGPT Atlas, in which the chatbot responded, “Trust No AI,” when a user asked for an analysis of a Google Docs file about horses that concealed malicious commands. Also, last month, a researcher tipped Microsoft off to a prompt injection vulnerability in Copilot that may have allowed attackers to trick the chatbot into stealing sensitive data, including emails.

In an emailed statement, Microsoft said its security team continuously tries hacking Copilot to find any prompt injection vulnerabilities, blocks users who try to exploit any found and monitors for abnormal chatbot behavior, among other tactics.

“Microsoft ensures its generative AI systems remain resilient against evolving threats for all our customers, including defense and national security,” the statement said.

Responding publicly to criticism on X, Dane Stuckey, OpenAI’s chief information security officer, wrote that “prompt injection remains a frontier, unsolved security problem, and our adversaries will spend significant time and resources to find ways to make ChatGPT agent fall for these attacks.”

Along the same lines, Caspi said, “You cannot prevent the prompt injection [fully], but you need to limit the impact.” He advised that organizations limit an AI assistant’s access to sensitive data and limit the user’s access to other organizational data.

For instance, the Army has awarded contracts worth at least $11 million to deploy Ask Sage, a tool that lets users restrict which Army data Microsoft Azure OpenAI, Gemini and other AI models can access to run queries and tasks. Ask Sage also isolates Army data from user prompts and external data sources.

Caspi, who is not an Army contractor, likened a prompt injection attack against an organization running Ask Sage to a lockdown situation where “you’ve got this insider, but it’s sitting in one room, and it can’t leave the room or carry out sensitive information.”

Andre Slonopas, a Virginia Army National Guard member and former Army cyber and information operations officer, uses Ask Sage and voiced confidence in the Army’s defensive AI tools, if not those of nuclear power plants or manufacturing entities, largely in rural, poorer areas.

The Virginia National Guard joined with essential services, such as power utilities, to help defend their networks against AI-powered cyberattacks, as part of a September simulation, given that service disruptions can jeopardize military preparations.

Typically, an adversary encrypts its network traffic to evade detection, but, for the sake of an experiment, organizers did not encrypt the AI offender’s traffic because “we wanted the blue team [of humans] to see exactly what the AI was doing,” Slonopas said.

“The blue team was absolutely defeated,” despite being able to watch the AI scanning its networks, creating fake usernames to gain unauthorized access and executing instructions to defeat the team’s systems.

“Whether the AI is doing prompt injection, spoofing or maybe even some sort of a brute force attack, the speed of AI is so unbelievably immense that simply human beings cannot counter it,” and, therefore, “you have to make cybersecurity AI more accessible and more affordable,” Slonopas said.

“If a water utility has to pay, say, $30,000 for a defensive AI license, well, it will amplify one person to be like 40″ or dozens of personnel, he said.

In response to questions, Army Cyber Command spokesperson Kyle Alvarez said in an emailed statement, “Due to the current lapse in appropriations, ARCYBER was unable to accept or respond to any media engagements or requests.”

Army contractors, too, are under attack from state-affiliated AI.

“China is using offensive AI like nobody else,” said Nicolas Chaillan, the founder of Ask Sage and a former U.S. Air Force and Space Force chief software officer.

“We see so many attacks coming after us,” all of which the company has stopped, Chaillan added.

A military official, who spoke on condition of anonymity due to the geopolitical sensitivity of the matter, said that China does “appear” to be the most skilled in offensive AI. However, the official added, AI spoofing and translation allow the United States, China, Iran, other countries, hacktivists and financial cybercriminals to masquerade as one another.

For example, the official said, “Right now, with ChatGPT, I can program in Chinese. I don’t speak Chinese, but because of the ChatGPT capabilities that I have, I can do that.”

Aliya Sternstein, J.D., is an investigative journalist who has covered technology, cognition, and national security since Napster shut down, working for various outlets including Atlantic Media, Christian Science Monitor, Daily Beast, Forbes Magazine and Just Security. She is also a research analyst at Georgetown Law.

In 2021 the 1st Multidomain Task Force, 5th Battalion, 3rd Field Artillery Regiment, 17th Field Artillery Brigade at Joint Base Lewis-McChord received all equipment — except for the actual live rounds — for the Long-Range Hypersonic Weapon capability it calls Dark Eagle.

The Washington state-based unit was supposed to get the missiles in fall 2023 but several aborted test events forced the Army to push back its fielding plans.

The first three missiles were distributed to the unit earlier this year, with the last of the munitions arriving in July. The fourth round is currently going through acceptance checkouts at Lockheed Martin’s facility in Cortland, Alabama, Lozano said. The final eight are expected to be delivered “by the end of December,” Lozano said.

The Army is also aiming to conduct a hypersonic missile test around the same time.

“That will complete the basic load for battery one, up at JBLM,” he added. “Then … we’ll start working on Battery Two.”

The fielding of the munitions follows a lengthy delay in the testing process. Several tests of the round were aborted in 2023 due to challenges at the range and the hiccups in the process of firing up the missile for launch.

Testing the all-up round was considered critical to ensure the system was safe, effective and ready for fielding.

The U.S. is presently in a race to field the capability and develop systems to defend against hypersonic weapons, which China and Russia are actively developing and testing.

The Army conducted an end-to-end successful flight test of its hypersonic missile at the Pacific Missile Range Facility in Hawaii in May 2024, which put the initial fielding to the first unit closer on the horizon.

While the process of equipping the Army with the weapon has taken nearly two years longer than planned, service officials have been quick to point out that missile development programs typically take about 10 years. The LRHW program, meanwhile, is only just beyond the five-year mark.

The Army has worked with Leidos’ Dynetics for years to build the industrial base for the Common-Hypersonic Glide Body that will be used by both the ground service and the Navy, as the domestic private sector has never built a hypersonic weapon.

The service also separately produced launchers, trucks, trailers and the battle operations center necessary to put together the first weapon battery.

Lockheed Martin is the weapon system integrator for the Army’s hypersonic capability that will be launched from a mobile truck.

In preparation for receiving the all-up rounds, the Army completed its delivery of the first hypersonic weapon capability — minus the rounds — to the Multi-Domain Task Force unit at JBLM two days ahead of its end-of-fiscal 2021 fielding deadline. The unit has been training on the system since the delivery.

The next phase of production will ramp up quickly once delivery to the first battery is complete.

“The second tranche of production is both Battery Two and the Battery One reload,” Lozano said. “That’s a very complex missile system. Although we’ve gotten through the testing of the system and we’re confident that it works, we’re still working through how to take a prototype system that’s complex and transition it into production, where you have a very repeatable process that drives out quality defects.

“We’re in the middle of that process right now,” Lozano added. “Things that we’re wrestling with every day is we’ll build up part of the missile and we’ll take the boost phase rocket motor and mate it to the missile midsection and then the inspector will come by and we’ll find some defects that we will have to work and address. … We’re learning as we build these missiles.”

Speed, of course, brings added challenges, Lozano acknowledged.

“Just the nature of having gone so fast — if you were doing a typical 10- or 15-year missile program, you drive all this risk out of the program and you’d be delivering on time. You’d just be delivering 10 years from now.”

Yet, “the good news is that we’re making progress now,” he said. “We just have to keep our foot on the gas pedal and continue to push forward and make more progress.”

The system is the Greyshark from Germany’s EuroAtlas, an autonomous underwater vehicle measuring 6.5 meters in length and weighing as much as a delivery van. It is one of the naval drones the German Navy brought to the Portuguese coastal town of Sesimbra to partake in NATO’s Dynamic Messenger and REPMUS exercises.

These drills, which focused on testing over 200 unmanned systems, ran for three weeks in local waters.

An important portion of the training was dedicated to mine countermeasures warfare and the protection of critical underwater infrastructure. European navies are in the process of retooling their naval de-mining procedures with the help of autonomous systems and forward-deployed sensors to keep human divers away from the dangerous work.

Exercise participants employed different types of naval drones to perfect mine hunting operations, which entails the complex work or coordinating the work of multiple robots simultaneously.

“Command and control, for instance, is much more complicated than it was before, because you’re bringing more [robotic] systems into the network – it requires more coordination, more people and a bigger orchestration of small tasks,” Cmdr. Andreas Montag, head of the maritime unmanned systems 3rd Minesweeping Squadron at the German Navy, said.

During REPMUS, the Greyshark, equipped with four high-resolution sonars, LIDAR scanners, electromagnetic detection and two camera systems, was tested for hunting submarines and searching for decoy mines. Officials said that it can reach ranges of up to 8,000 nautical miles, or from Sesimbra to New York and back.

The German Navy also demonstrated two other naval drones from vendor EvoLogics: the Sonobot unmannedsurface vessel and the Quadroin AUV.

The Sonobot was described as acting as a “surface gateway” throughout the exercise, providing underwater positioning and data networks – both wifi and GPS underwater – to other autonomous systems. It is also in use with the Ukrainian military to survey water bodies and detect dangerous objects.

The Quadroin is the underwater variant of the Sonobot, and up to six were used simultaneously during certain missions to look for mines as a swarm.

“For the first time this year, we were able to not only detect mines but also identify them and get a video picture from each one during a single mission – for this the Quadroins split up their roles, some ran a sweep search of the sea floor while others directly swarm to the target and took a picture,” Rui Madeira, commander of the divers group of the Portuguese Navy said during a demonstration on Sept 23.

From manual to machine

Each week, the scenarios and tasks assigned increased in difficulty to test the readiness of the different units. For example, diver groups initially knew the zone they had to work in, but would later on find out only shortly before their missions and faced strict time constraints to complete the objective.

The French Navy tested the A9-M AUV, manufactured by Exail, during naval mine warfare operations. The platform was first deployed to survey the seabed, the sonar data were then evaluated on shore, and the potential mines were confirmed by divers or remotely operated vehicles.

For Louis, an officer with the French EOD diver group, the next phase regarding the use of unmanned technology in mine warfare operations will be the integration of artificial intelligence.

“Integrating AI into these drones would allow them to independently classify the threats they detect in the field – currently, the majority of the classification of sonar data is done by human operators after they’ve extracted the data from the system and analyzed on a computer onshore,” the officer, who withheld his last name for security reasons, told Defense News.

While technology is progressing quickly, some hiccups remain.

This was the case with the British Navy, which relied on the REMUS 100 AUV from the company HII to carry out collaborative missions with the U.S. and Dutch Navies.

One naval officer said that on a few occasions during the training, the vehicle overheated due to high temperatures experienced earlier this month. Ad-hoc solutions included keeping the drone cool with blankets and deploying it earlier in the day.

Another mishap involved two Sonobots from the German Navy that collided during a demonstration. Plus, at one point, a Portuguese fishing boat ventured into the exercise zone and struck the French A-9M system, damaging the drone enough that officials had to take it out of the demonstration.