The navy is experimenting with its simulator drills to put crews in “increasingly stressful situations,” said Capt. Jérôme Henry, the head of training for the navy’s surface personnel, at the Paris Naval Conference this week. Henry said he’s drawing on past experience as commander of the frigate Alsace, which came under attack multiple times in the Red Sea, to “toughen up our crews.”

“What I saw in the Red Sea is that when you’re under intense stress, people react more or less well but in any case, you lose some of your composure, you get what’s called tunnel vision,” Henry told Defense News on the sidelines of the conference. “If we’re going to be in high-intensity combat, our crews need to be ready for that stress, and the question is, how are we going to prepare them?”

French, UK naval chiefs urge dramatic changes in warship design

Training tweaks include crews going for a run or doing push-ups right before stepping into weapon simulators to get heart rates up, creating sensory overload by adding noise, smoke and drone swarms to simulations, and adding weapon malfunctions in drills, said Henry, who took on his current role last year.

Henry says he adopted the idea of stress drills from the French Navy’s special forces, the Commandos Marine, and is seeking to find out how American and Israeli forces include stress in their training.

The goal for now is to dial up stress levels “as high as possible” to ensure that reflex actions are always the right ones, the training chief said. Henry said the challenge is the difficulty of putting people under such stress “that they feel like their final hour has come.”

The training division is trying to create the most disruptive environment possible in its simulators, so that personnel including gunners and missile operators “can mechanize their actions” to ensure they’ll be able to perform in combat whatever the situation, according to Henry.

“I know what it feels like to take a missile on the nose at four times the speed of sound,” Henry said. “So we know where we start to get stressed, we know we have to prepare for that.”

The most important lesson learned from the Red Sea is the need to be ready at all times, Adm. Harold Liebregs, commander of the Royal Netherlands Navy, told Defense News.

“The time when we could leave port, then build up and then see what mission we were going to do, that has changed,” Liebregs said. “It’s about training, but also about buildup. It’s about everything becoming more and more realistic, and it starts with having your war plans ready.”

Liebregs said the officer who commanded the support ship Karel Doorman during its deployment in the Red Sea, Paul Bijleveld, is now the navy’s commander for Sea Training, “so all the lessons we learned there, he’ll take onboard. Perhaps that is no coincidence.”

Western navies lack combat experiences in terms of high-intensity naval war, said Capt. Bryan McCavour, deputy assistant chief of staff for information warfare at the United Kingdom’s Royal Navy, who spoke on a panel with Henry. With fewer and fewer platforms in national fleets, ongoing problems with ship maintenance and availability, training is consistently getting compressed, he said.

“If we’re going to engender that war-fighting spirit and maintain it, and have that culture as a decisive factor in battle, we need to invest more time in high-end war-fighting training than I think we currently do,” McCavour said.

McCavour said it’s been longer since the Falklands War than between that conflict and World War II, and “combat-ready naval forces in that sense maybe don’t exist today in the way we think, because it’s been a very long time since we had a high-end conflict.”

He said Russia was reminded of that lesson with the sinking of the cruiser Moskva in the Black Sea, and Western forces need to take that into account when they look at responding in the South China Sea or the High North around the Kola Peninsula.

The Red Sea was also “a little bit the rediscovery” of low-end threats, with renewed focus on small-caliber weapons and cannons, and a pipeline of defensive layers including jamming, light missiles and laser-guided rockets, according to Henry, who directed the navy’s annual Wildfire drone exercise to focus on saturation as well as the risk of friendly fire in a busy environment.

“When you have a lot of things flying around you, and you open fire with other friendly units, we saw in the Red Sea that mistakes can be made,” Henry said. “So we’re working on that.”

The French Navy is furthermore training for combat while minimizing radio emissions, relying on adversary emissions to build situational awareness, according to Henry. The force is working to cut reliance on satellite positioning, helped by improved inertial-navigation systems as well as astral sights using the stars for positioning.

Henry mentioned the ‘Back to the ‘80s’ exercises by the French carrier strike group, which entails forgoing satellite communications and instead using HF and UHF radio for comms, “and above all, to be more frugal in our exchanges.”

Cyber warfare is the prime example of a threat likely to deprive naval forces of capabilities “at the worst possible moment,” according to Capt. Florian El-Ahdab, commanding officer of the French frigate Languedoc. He said preparing for eventualities such as loss of connectivity requires ‘Back to 80s’ type exercises and placing forces in situations of “great discomfort.”

“Today’s sailors reflect today’s society, so if I took your smartphone away today and told you to go somewhere, I’m not sure you’d feel very comfortable,” said El-Ahdab. He said “it’s the same thing” for the navy.

“If you take away all the tools available to the commander today, and all the amazing tools that are currently being developed, if I suddenly tell you that all of that is no longer available for one reason or another, how would you respond?” El-Ahdab asked. “That seems like a very, very good challenge to explore.”

The Singapore Airshow, organized here from Feb 3-8, marks the first time the T-Scope has been shown overseas. The device was designed specifically for unguided munitions and uses a variety of sensors and AI to instantly calculate and predict the projectile flight paths, taking into account environmental and physical factors.

“What we’re doing rather than placing sensors on the rocket itself, is we are putting them into an AI-driven guidance kit, which essentially transforms unguided rockets into precision-guided munitions to improve their accuracy for cheaper,” Alan Kuo, Tron Future’s director of sales, told Defense News.

Taiwan has been gradually building up its inventory of anti-tank weapons, including an unknown number of unguided rocket-propelled grenades, to deter potential attacks.

Wang said they were approached by one of the country’s leading defense institutes to develop a device that could help quickly mobilize personnel, if needed.

“By combining automated complex ballistic calculations and displaying a live corrected aiming point, T.Scope allows soldiers to achieve marksman-level proficiency within a very compressed training timeframe,” Kuo added. “It would allow for combat readiness even under limited preparation.”

Although it has already been tested, the company is hopeful that it will receive army certification before the end of the year.

Further testing has already been planned to see how it performs in shallow water, per a request from the undisclosed defense institute.

The National Chung-Shan Institute of Science Technology, or NCSIST, is largely regarded as the leading research and development center in the defense sector in Taiwan. It focuses on designing and producing indigenous weapons, including missiles and unmanned systems.

Tron Future’s counter-drone radars have been used by the Taiwanese Army for several years. The company has also been assisting the government in integrating counter-drone technologies with Taiwan’s low Earth orbit satellites through its T.SpaceRouter user terminals to increase the island’s communication resilience.

Col. Corey Klopstein, program executive officer for Operational Test and Training Infrastructure and commander of System Delta 81, described the program’s focus on realism and readiness during a media roundtable at Space Industry Days in Los Angeles, California, on Friday.

SWORD, the Space Force’s primary synthetic training environment, is a cloud-enabled, digital simulation platform designed to replicate contested space operations, including orbital dynamics, electronic warfare, cyber effects and adversary tactics. It allows guardians to train in realistic scenarios without relying solely on live, on-orbit assets or centralized facilities.

The platform has been demonstrated in large-scale exercises like Space Flag, supporting hundreds of guardians in realistic training, and is being scaled for broader enterprise use.

Klopstein stressed that SWORD is being developed to deliver the highest possible realism and adaptability.

“Our intent is to develop SWORD and make it as realistic as possible and increase the fidelity as we work closely with our users to understand what they need … and increase that work with our contractors to increase the fidelity of SWORD,” he told reporters.

That realism is achieved through rigorous validation. Digital models in the SWORD training environment are continuously being updated and refined by cross-checking against hardware-in-the-loop facilities (integrating real components into simulated environments) and live on-orbit assets.

“It’s a constant back and forth in trying to increase the fidelity of your digital environment and make it as realistic as possible,” Klopstein said, stressing that the Space Force cannot rely on synthetic data alone.

Speed is equally critical. With adversary tactics evolving rapidly, Operational Test and Training Infrastructure is prioritizing rapid integration of new threats into SWORD.

Klopstein highlighted close coordination with intelligence elements and the National Space Intelligence Center to ensure timely updates to red threat emulations.

“We’re working closely with our S2 and the Field COM S2s, as well as NSIC, to get the latest information that we can and leverage that information to provide updates to any red threat emulations within SWORD … to ensure the greatest accuracy possible,” he said.

These efforts support a long-term vision of enterprise-wide access, transitioning SWORD to a cloud-based infrastructure so guardians can conduct realistic, distributed training from home stations rather than centralized facilities.

While current implementations are site-specific, Klopstein described the goal as creating “backyard ranges” so guardians can “train from their home station, using the synthetic environment as the source of the truth.”

The SWORD program is supported by 10-12-month agile acquisition cycles to close near-term training gaps, alongside a planned fiscal 2027 cloud pathfinder, as reported by Air & Space Forces Magazine and Breaking Defense.

The $1.2 million project adds to the Pentagon’s growing investment in the $3 billion market for brain-computer interfaces (BCIs) — helmets, earbuds, other wearable devices and medical implants that use artificial intelligence to make sense of brain signals.

Elon Musk’s Neuralink, perhaps the most high-profile implantable example, has enabled clinical trial patients who cannot move or speak to type words that they imagine. Since at least the 1970s, the Defense Department has funded the development of BCIs that, among other things, restore eyesight and control prosthetic arms.

Now, advances in AI enable BCIs to interpret or change brain activity with unprecedented accuracy and speed. So much so that some scientists and neuroethicists say the capacity to monitor signals that reveal a warfighter’s or a veteran’s mental states, visual perceptions or inner dialogue may outpace the ability to shield that data from misuse.

“One could certainly imagine how enforced use of such devices could create a very dystopian basis for behavioral control,” warned James Giordano, director of the National Defense University’s Center for Disruptive Technology and Future Warfare and Georgetown University Medical Center’s former chief of neuroethics studies.

Consequently, any accessing or use of neural data must require “ongoing, active informed consent,” so active and retired military have the chance to opt in or out without penalty, he said.

Giordano, who was sharing his personal views and not speaking on behalf of any government agency, added, “The security of the neurological data used for that individual — and in that individual’s wellbeing only — is paramount. It goes beyond simple HIPAA.”

He was referring to the 1996 Health Insurance Portability and Accountability Act, a law regulating healthcare providers that predates global Wi-Fi and does not cover most app and algorithm developers, commercial cloud providers or consumer wearable designers.

Neurable and Air Force officials say that the non-invasive cognitive fitness tracker will build upon technology in the firm’s consumer BCI product — Master & Dynamic luxury headphones containing fabric electroencephalogram, or EEG, sensors and electrically conductive inks to gauge mental focus.

The $500 device, which launched last June, deploys an algorithm that maps brainwaves to ground-truth markers of focus — users’ speed and accuracy in finishing tasks, while undistracted and distracted. When the wearer’s brainwaves resemble those tied to flagging attention, the headphones sync with an app that alerts the user: “You’ve earned a Brain Break.”

The Air Force-funded BCI project will use other headset models, such as noise protectors or helmets, said Neurable co-founder Adam Molnar, who helped secure $4 million for prior military research and development work.

The study will measure cognitive performance with and without pressure — for instance, memory skill with and without sleep deprivation — against brainwaves to identify brain activity associated with optimal and suboptimal performance, according to Neurable.

The idea is that alerting service members to significant brain signal changes throughout the day — or, providing “neurofeedback” — will train them to adjust their behavior and brainwaves to hit peak performance.

“This is all within this concept of readiness. How do we make sure that we have a ready and able defense?” said Molnar, who anticipates wrapping development within two years.

Molnar added that the Space Force has a “really cool program where they’re using the Garmin watch,” embedded with heart rate, blood oxygen and other wrist-worn sensors, “to help improve cardio fitness and physical activity‚” but “we don’t really have that for the brain.”

A comparable neurofeedback tool, he said, may spur “more robust brain health practices, not just for defense, but also for possibly identifying Parkinson’s or Alzheimer’s disease indicators a decade before you feel your first symptom.”

Air Force Research Laboratory officials said that Neurable’s fabric-based sensing technology may aid multiple operations, inside and outside of the cockpit.

“I suspect that there are a large number of military members for whom a real-time fatigue or cognitive state monitor would be appealing if they thought it would make them more effective at executing the mission,” William Aue, cognitive neuroscience section chief at the Air Force Research Laboratory’s 711th Human Performance Wing, said by email.

Let them ‘make their mind up’

Some neuroethicists, however, cautioned that the use of brain wearables may also invade troops’ privacy and foster prejudice.

“We need to be much more reflective on, much more concerned about and much more focused on the implications of altering a person’s brainwaves,” through neurofeedback training, “as compared to having them work out more or do more pull-ups,” said Jared Genser, a lawyer and co-founder of the Neurorights Foundation, a group advocating for the legal protection of people’s brain data.

Similarly, Giordano said, “The concept of personalized medicine must be applied here,” meaning that military leaders must tailor training and performance metrics to the limits of each individual’s brain and body, so “that you don’t get plateaus, and you’re not beginning to see performance decrement or performance fatigue.”

The Neurorights Foundation’s Medical Director Sean Pauzauskie, a neurologist in the UCHealth University of Colorado system, added that cognitive fitness standards may overstep the proposed right to freedom from algorithmic bias based on neural data interpretations.

“If the warfighter’s brain is incapable of the appropriate plasticity to achieve the desired [mental] state,” through neurofeedback training, “they could be discriminated” against by their superiors, Pauzauskie said.

Also, EEG data flowing into and out of consumer wearables can reveal several diseases, including epilepsy and mild cognitive impairment, as well as a range of emotions.

Via email, Molnar responded to such concerns, saying, “Neurable’s focus is on supporting human performance and safety, not enforcing neurological conformity.”

“Any feedback provided by our systems is designed to be informational and user-centered, helping individuals understand factors like cognitive workload or fatigue,” he added, “much like a heart-rate monitor provides insight into physical exertion.”

Neurable officials said the system’s sensors filter out all EEG data not needed by its algorithms and encrypt the limited amount of necessary data in transit.

“Even if someone hacked someone’s brain data, they would not be able to make sense of it,” Molnar said.

The Air Force’s Aue confirmed that, during experiments, the Air Force lab anonymizes EEG data streaming from Neurable’s wearables, adding that the information is “challenging to interpret without the algorithms we use to process the raw data.”

But, such protections may not stop adversaries from decoding wearable-based intelligence in the future, as AI becomes smarter and devices utilize more EEG data.

Already, in the laboratory, algorithms can narrow down a person’s identity into a “mindprint” of their mental processing or “brainprint” of their brain structures, based on EEG data from a medical headcap, Pauzauskie noted.

Likewise, Genser explained, EEG data that is out into the open, if isolated from a person’s name and identity, will remain deidentified but only for a period. The problem is that, as generative AI-powered software develops, “you are going to be able to decode a lot more information from the data that escaped, enabling the data to be reidentified and traced back to you personally.”

In an email, Molnar acknowledged that as “AI continues to accelerate research and development, it is prudent to anticipate how capabilities might evolve and to establish safeguards early.”

He suggested additional ethical review, clear boundaries on acceptable uses, and dialogue across technologists, ethicists, policymakers and technology users.

Aue, who has been working on other research with Neurable for four years, said in an email that research study participants “are fully informed about the nature of the work before participating.”

That said, as the technology matures, conversations “need to occur about the ethical use of the technology and what military doctrine might entail,” he added.

Giordano said that the key to any adoption of AI-based neurotech in the military is mandatory informed consent during and after service.

Personnel must receive updates on any adverse effects and information on “the availability or non-availability of continuity of care if things go wrong,” he added. “That way, they can make their mind up.”

Reading the minds of wounded warriors?

Due to AI, implantable brain devices, much as non-invasive brain wearables, now offer troops and veterans novel insights into themselves — and possibly the same intel to adversaries, too.

For instance, new eyesight neuroprostheses rely on AI to “bridge artificial vision and neural tissue,” said Christopher Steele, chief of strategy at the Pentagon-backed nonprofit Medical Technology Enterprise Consortium. The consortium provided $2 million to help develop a prototype.

The camera-equipped BCI uses AI to decode brain signals that convey the landscape a person’s eyes are trying to see. AI then translates data from the camera and other external sensors focused on that view into artificial signals that the mind understands as shapes and motion.

Finally, AI compares the desired and perceived scenes and adjusts the artificial signals to show more detail, said Steele, a former director of the Army Medical Research and Development Command.

This self-adjusting AI, while a potential breakthrough for visually-impaired warfighters, also introduces risks, Giordano said.

“The AI that provides continuous closed-loop signals between the eye and the brain is not monitored,” he explained.

As such, it may self-adapt in a way that falsifies images, omits certain elements of the view or delays transmission of an image, any of which may result in a loss of situational awareness, Giordano said.

Compounding the problem, when the AI “system itself is vulnerable to hacking,” an adversary can monitor or corrupt the warfighter’s visual input, he added.

“If we deploy this device downrange” in a combat zone, ”can it be hacked by an enemy?" Steele said. “We are concerned about the hacking of any medical capability.”

Currently, the system only operates in clinical trials that are subject to HIPAA, and other protections include hardware isolation, encryption and AI guardrails that prevent unsafe adaptation, he added.

Yet, as Giordano cautioned, self-learning AI, by design, can learn good or bad behavior that may override guardrails.

Another military funding arm, the Amyotrophic Lateral Sclerosis Research Program, is underwriting clinical trials to evaluate Braingate, a forerunner to Musk’s Neuralink implant that translates words a patient imagines into audible speech.

The Defense Department’s $2.3 million contribution aims to help people with ALS — a disease associated with military service — who cannot move or talk.

This advance is “foundationally fantastic” for people with various neuro-motor conditions who can think of speech but not generate motor output to move their mouths, meaning “they’re communicatively locked in,” Giordano said. The device “allows these individuals the autonomy of communication.”

At the same time, because the device reads signals in the motor cortex that show attempts at moving the mouth, the tool allows for “what might be regarded as mind-reading,” Giordano said.

Currently, human beings, military or not, have no legal right to be free of mind-reading.

Neither international human rights law nor federal regulations safeguard mental privacy, mental autonomy (“free will”) or neural data, according to the Neurorights Foundation.

Only four states — California, Colorado, Connecticut and Montana afford protection to “neural data” within their state consumer data privacy laws. And one country, Chile, has adopted a constitutional amendment to protect “mental integrity.”

As the Foundation’s Genser notes, one of the only countries with a national-level set of principles on the ethical use of BCIs, China, may be the least likely to use it.

He is pressing for “clear limits” on the purposes for developing and using neurotech and neural data in the military, subject to strong oversight that safeguards the health and autonomy of service members.

“If we don’t have any tools or standards around the national security implications of emerging neurotechnologies, then how is our military different than any other military?” Genser posited.

As an illustration of unregulated brain alteration, he pointed to the CIA’s MKUltra program, a Cold War-era mind-control experiment that tested LSD and other psychedelic drugs on soldiers and others, sometimes with permanent psychological and physical harm.

“The last thing we want is for neurotech to be the next version of MKUltra,” Genser said, “not using drugs, but using neuro-stimulation or other neural approaches to alter brain activity.”

Clive Higgins, the CEO of Leonardo UK, said that online bot activity promoting the independence of Scotland from the UK reduced as Iranian authorities violently suppressed civil protest this month.

The implication is that Iran has been pushing the cause of Scottish independence through malign cyber activity in a bid to weaken the UK, but slowed its activity as violent protests raged in Iran.

“What has been very interesting over the recent period, given the social unrest in Iran, is the drop off in social media activity and chat bot activity focused on Scotland, elections and independence,” Higgins told Defense News.

On Sunday, Scotland’s first minister John Swinney said he would call for a referendum on Scottish independence if his SNP party wins a majority in Scottish parliamentary elections in May. At a referendum on independence held in 2014, 55% of Scots voted to stay in the U.K.

A report produced by the British parliament’s intelligence and security committee last July said the U.K. was a priority target for Iranian cyber attacks, ranking just behind the U.S. and Saudi Arabia.

“Iran has a high appetite for risk when conducting offensive activity and its intelligence services are ferociously well-resourced,” said committee chairman Kevan Jones at the time. He added, “It supplements this with its use of proxy groups – including criminal networks, militant and terrorist organizations, and private cyber actors – to provide it with a deniable means of attacking its adversaries with minimal risk of retaliation.”

Leonardo has been raising its profile as a cyber security provider, and Higgins said, “We are very clearly aware of the increasing threat from a cyber perspective and our capabilities monitor very regularly state actors who are mobilizing chat bots and social media accounts to disrupt the normal way of life we all enjoy.”

He added, “We are able to see very clearly the actions of state actors where we can identify IP addresses and other elements that articulate to us where things are being done, what kind of messages they are trying to deliver and how they are trying to disrupt U.K. society.”

He said, “There are three or four main state actors. We see a lot from North Korea, Iran, Russia and China. I don’t want to say that is state sponsored activity. It is where the activity derives from.”

“Very soon we will have the world’s leading AI models on every unclassified and classified network throughout our department,” Hegseth said in a speech at Musk’s space flight company, SpaceX, in South Texas.

The announcement comes just days after Grok — which is embedded into X, the social media network owned by Musk — drew global outcry and scrutiny for generating highly sexualized deepfake images of people without their consent.

Malaysia and Indonesia have blocked Grok, while the U.K.’s independent online safety watchdog announced an investigation Monday. Grok has limited image generation and editing to paying users.

Hegseth said Grok will go live inside the Defense Department later this month and announced that he would “make all appropriate data” from the military’s IT systems available for “AI exploitation.” He also said data from intelligence databases would be fed into AI systems.

Hegseth’s aggressive push to embrace the still-developing technology stands in contrast to the Biden administration, which, while pushing federal agencies to come up with policies and uses for AI, was also wary of misuse.

Officials said rules were needed to ensure that the technology, which could be harnessed for mass surveillance, cyberattacks or even lethal autonomous devices, was being used responsibly.

The Biden administration enacted a framework in late 2024 that directed national security agencies to expand their use of the most advanced AI systems but prohibited certain uses, such as applications that would violate constitutionally protected civil rights or any system that would automate the deployment of nuclear weapons. It is unclear if those prohibitions are still in place under the Trump administration.

During his speech, Hegseth spoke of the need to streamline and speed up technological innovations within the military, saying, “We need innovation to come from anywhere and evolve with speed and purpose.”

He noted that the Pentagon possesses “combat-proven operational data from two decades of military and intelligence operations.”

“AI is only as good as the data that it receives, and we’re going to make sure that it’s there,” Hegseth said.

The defense secretary said he wants AI systems within the Pentagon to be responsible, though he went on to say he was shrugging off any AI models “that won’t allow you to fight wars.”

Hegseth said his vision for military AI systems means that they operate “without ideological constraints that limit lawful military applications,” before adding that the Pentagon’s “AI will not be woke.”

Musk developed and pitched Grok as an alternative to what he called “woke AI” interactions from rival chatbots like Google’s Gemini or OpenAI’s ChatGPT. In July, Grok also caused controversy after it appeared to make antisemitic comments that praised Adolf Hitler and shared several antisemitic posts.

The Pentagon did not immediately respond to questions about the issues with Grok.

The new specialty, designated 49B, will be open to eligible officers through the Voluntary Transfer Incentive Program, beginning in January 2026, with the goal of developing a group of leaders that “advances the Army’s ongoing transformation into a data-centric and AI-enabled force,” the statement said.

The Army said officers with applicable backgrounds would use their expertise to accelerate decision-making, sharpen targeting, improve logistics efficiency and support the fielding of robotics systems on the battlefield.

“This is a deliberate and crucial step in keeping pace with present and future operational requirements,” Lt. Col. Orlando Howard, an Army spokesperson, said in a statement. “We’re building a dedicated cadre of in-house experts who will be at the forefront of integrating AI and machine learning across our warfighting functions.”

The statement said the program was open to all officers eligible to voluntarily transfer branches, but those with related background experience or education “will be particularly competitive candidates.”

The officers chosen for the new pathway will get graduate-level training and hands-on experience with AI-enabled systems, the statement said, adding that officers will be reclassified by the end of fiscal year 2026.

The Army is considering broadening the program in the future to include warrant officers.

The new officer roles were announced shortly after the Pentagon launched GenAI.mil, an artificial intelligence platform for the Department of Defense, and then announced an upcoming expansion of available tools.

In 2025, the Army also introduced a Robotics Technician specialty for warrant officers, meant to provide brigade and special forces formations with information on robotics, AI and machine-learning.

The report, released on Nov. 17, highlights security lapses across 10 DOD components linked to digital activity from personal and government devices, online communications and defense platforms that “generate volumes of traceable data.”

“Massive amounts of traceable data about military personnel and operations now exist due to the digital revolution,” according to the report, which highlighted ship and aircraft movements as examples of accessible information. “Public accessibility of this data enables malicious actors to exploit critical information and jeopardize DOD’s mission and the safety of its personnel.”

The report found that U.S. Cyber Command, the National Security Agency, the Defense Intelligence Agency, the Defense Counterintelligence and Security Agency, U.S. Special Operations Command and every branch of the military failed to adequately address two critical areas — training and security assessments — for reducing the risk of digital threats.

Of the components assessed, only U.S. Special Operations Command was found to have consistently trained personnel about the risks of digital information in the public across all relevant security areas.

The report also found that eight of the 10 DOD components did not conduct threat assessments across the required security areas of force protection, insider threat, mission assurance and operations security.

The GAO goes on to illustrate how threat actors can steal public digital information about DOD operations and personnel, thereby posing operational and national security risks.

Information transmitted through press releases, news sources, online activity, social media posts and ship coordinates could, in theory, be used to project a vessel’s route and disrupt an aircraft carrier’s operations, the report states.

The GAO noted that three of five offices under the Office of the Secretary of Defense have issued policies and guidance on risks associated with public access to digital information.

Those initiatives, however, are too “narrowly focused” and do not include all relevant security areas, the report states.

The Defense Security Enterprise Executive Committee is “well-positioned to lead a department-wide collaborative assessment of policies and guidance on digital footprint and profile risks,” according to the report’s authors.

“Without such an assessment, DOD will have difficulty in determining whether risks are being sufficiently managed within the boundaries of their legal authorities,” the report adds. “Also, DOD will face ever-increasing threats to personnel privacy and safety, mission success, and national security.”​

The GAO report made 12 recommendations to DOD to assess its policies and guidance, collaborate to reduce risks, provide training on the digital environment along with its associated risks across security areas and complete required security assessments.

DOD agreed with 11 of 12 recommendations and partially concurred with one.

Among the recommendations in the GAO report are instructions for the service secretaries of the Army, Navy and Air Force to conduct “required assessments in the security areas of force protection, insider threat and mission assurance.”

The report also recommends that Defense Secretary Pete Hegseth oversee the implementation of security assessments for the other listed components.

On Capitol Hill, Congress has echoed similar security concerns targeting the Defense Department. In an effort to mitigate national security risks, both the Senate and the House have introduced legislation in recent years with recommended provisions to protect the digital footprint of DOD personnel.

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.

But if you need JTACs to call in air strikes, and space-based weapons are becoming a reality, then don’t you also need specialists to call in space strikes? That’s why there should be Space Joint Terminal Attack Controllers, or SJTACs, argue two experts.

For example, “the space JTAC would connect tactical, on-the-ground SOF [special operations forces] units with space assets for targeting adversary military airbases, critical infrastructure, and more complex targets, such as Russian floating nuclear power plants,” wrote retired U.S. Army colonel Kevin Stringer and Marius Kristiansen, a Norwegian Army officer, in a July essay for the Irregular Warfare Initiative website.

Unlike space personnel already assigned theater special operations commands, SJTACs would embed with special operations tactical units, the essay suggests. They would “enable the assessment of vulnerabilities, ensuring precision in any potential attack, as well as monitoring target activities, tracking movements, and providing real-time situational awareness for preemptive strikes or future sabotage missions,” Stringer and Kristiansen wrote.

SJTACs are needed to coordinate the advent of space-based weapons that paint a future where orbital bombardment of ground targets is the norm, the pair argues.

“As space capabilities develop from science fiction to reality, the SJTAC could access future space weaponry ranging from lasers, electromagnetic pulse (EMP) weapons, to the currently theoretical ‘Rods from God’ kinetic bombardment concept,” Stringer and Kristiansen wrote.

Just as JTACs are needed to target airstrikes and avoid hitting friendly forces, SJTACs would perform the same role for space.

“For normal airstrikes, an infantry scout or aviation aeroscout can call them in during an emergency,” Stringer told Defense News. “But the margin of error grows with a generalist doing infrequent specialist activities. A JTAC is a professional who focuses on this task. The SJTAC would be similar.”

There has long been an intimate link between space and special operations forces. The “Cyber-Space-SOF” triad is seen as a strategic capability, while U.S. Space Force is now standing up its own special operations component as part of U.S. Special Operations Command (SJTACs would belong to that Space Force component).

“Combined with space and cyber capabilities, SOF can access satellite communications, space-based reconnaissance, and cyber tools to disrupt enemy activities while maintaining a low signature,” the essay explained.

One question is redundancy. Rather than having separate SJTACs, would it be simpler to offer additional training in space capabilities to existing JTACs? Such training could be added to the Army-run Special Operations Terminal Attack Control Course, or SOTACC, which certifies JTACs. Or, the space role could be handed to Air Force combat controllers.

Stringer and Kristiansen disagree. In their view, SJTAC critics “overlook the fact that with space defined as a separate warfighting domain from the air, and with the creation of U.S. Space Force as a separate service, organizational logic and the development of deep space expertise necessitate a division of labor that would place the SJTAC function firmly within the Space Force sphere of responsibility.”

Separate STACs “would also avoid the inefficiency of having SOF from each service — Army, Navy, Air Force, and Marines — develop their own SJTAC capabilities," the essay argues.

SJTACs would also be useful for coordinating U.S. space capabilities with NATO and other allies, Stringer and Kristiansen wrote.

“Given that not all NATO countries have space assets, and the planning assumption is that NATO SOF formations will operate in a combined fashion, a space JTAC becomes a critical linkage for allied interoperability,” the essay suggested.

Ultimately, if the special operations community wants to play a role in space-based capabilities, it will need to have its own space specialists, Stringer argues.

“Without this function, SOF will not be able to access the developing space capabilities nor be involved in their development and experimentation,” Stringer told Defense News.

Part of the exchange, which matched up troops from U.S. Army Southern European Task Force, Africa, with personnel from the Angolan Military Health Division, sought to identify emerging, scalable tech capable of real-time language translation in what could one day be high-stress environments in which interpreters are not readily available.

As select personnel in the exchange trained on an array of medical procedures, others, like 1st Lt. Tucker Chase, a PAO assigned to SETAF-AF, worked on translations between U.S. personnel and their allies from Angola, a Portuguese-speaking nation in southern Africa that also boasts more than 40 local languages.

“Translation tools like these help us engage more directly with our counterparts and ensure their voices are accurately represented,” Chase said in a service press release. “We’re always looking for ways to improve how we tell the story of our missions.”

And improvements were indeed established. By the end of the exchange, which was conducted over the last week of September, soldiers had zeroed in on at least one translator app capable of processing smooth conversation by way of side-by-side — or face-to-face — displays without needing to hand the phone back and forth, according to the release.

Thanks to the technology’s hands-free, auto-translate function, communication between the allied partners continued to flow without interruption, the release said.

While not necessarily replacing human interpreters, that capability could prove vital in overcoming language barriers in the midst of what could be life-or-death situations.

“I would pull out my phone and open the app,” 1st Lt. Matthew Higgerson, a perioperative nurse with Landstuhl Regional Medical Center, said in the release. “We’d basically just talk regularly as if we’re having a conversation, then give each other time to read what was translated and continue talking.”

In another portion of the exchange, which was hosted on site in Angola’s capital city of Luanda, personnel were able to employ the translator tool during a joint medical press conference conducted by both nations.

Chase noted that the tech proved its value during that exercise, as well.

“It’s a practical solution when resources are limited,” he said. “It’s intuitive, fast and helps us maintain the authenticity of our engagements.”

Though it was brief, the exchange program marked one small part of an ongoing effort by SETAF-AF to stress-test military technology across the continent’s 54 sovereign nations, much of which is aimed assisting humanitarian emergencies.

IPPS-A active-duty administrators told reporters at the Association of the United States Army’s annual meeting on Tuesday they planned a “disruptive technology pilot” in the coming year. Known as the HR Intelligent Engagement Platform, it is designed to “push the envelope” on building smart tech into existing processes, Col. Matt Paul, project manager for IPPS-A, said.

“Were using the Army’s open solicitation and going through a competitive down-select process to do some prototyping, to determine, how can we use AI – we’ve got all these disparate HR and non-HR related systems out there in the wild – how can we use AI to bridge that?” he said.

“So, if I’m a soldier and I’m just on a prompt, how do I use a single prompt to get the information that I need?” he asked.

An agentic AI tool being evaluated could offer soldiers a help desk within the IPPS-A mobile app to troubleshoot issues and help locate functions, Paul said. In the future, he said, AI may take over some form-writing entirely, he said.

“Next level would be how to actually initiate HR transactions from a prompt,” he said. “I’m Col. Matthew Paul, and I need a leave form, and then it might ask me some questions, and then I might respond, and at the end of that conversation, I may have an approved leave form, all through AI prompting. I don’t know if we’re going to get there, but that’s the whole point of the pilot, to push the envelope, to see how far we can get.”

The IPPS-A team is looking to award a contract for the pilot program this year, Paul said, but does not have a fixed timeline for its execution.

“It could be three months. It could be eight months,” he said. “We’re trying to maximize flexibility and go where the technology leads us.”

IPPS-A, an investment of nearly $600 million for the Army, went live across the service in January 2023 after several years of rollout delays. While the system has prompted its share of user complaints, particularly in its early days, officials said they are continually pushing out updates and soliciting feedback as they absorb new processes from existing systems.

In May, IPPS-A took over creation of DD-214 discharge papers, allowing soldiers to see draft DD-214s on their mobile devices and correct errors before leaving the service.

“When we released it, it was a minimal viable product,” Col. Becky Lust, director of the Army’s Functional Management Division, said.

“Since May … I’m not sure there’s been a release that we haven’t added some functionality,” she said.

In coming months, IPPS-A officials are also looking to take over the process of checking soldiers into new units and detaching them from old ones, a task that would require multiple stops at physical locations for paperwork processing and approval. Within eight months, Paul and Lust said, they hope to launch a new, mobile-based workflow.

Soldiers “would be able to show up at the location and maybe scan a QR code that says that I’m here, and that will trigger the [Military Personnel Division] to initiate in-processing. And then the soldier will get what we call activity guides in IPPS-A, and that will guide him through each of the spots that they want him to visit,” Lust said. “But he visits them all virtually, so he gets the information about that installation, and he can click yes or no if there’s questions that they need. And then he doesn’t really have to go to multiple spots … And then he’s building readiness for the Army, because then he will be able to go to his unit much quicker.”

Also in the works, with the help of soldier officer, enlisted and warrant officer working groups assembled over the summer, are improvements to the soldier talent profile functionality on the app that will allow troops to put all their relevant experiences and skills in one place for future civilian job hunts. Updates on that project will be pushed out over the next 90 days, Lust said.

“The first shots will be fired in the cyber domain,” said Maj. Gen. Jake Kwon, director of strategic operations for the Army’s Office of the Deputy Chief of Staff. “The Army has to think differently, and we have to fight faster.”

Kwon spoke Tuesday alongside Lt. Gen. Jeth Rey, deputy chief of staff G-6, and Brandon Pugh, principal cyber advisor to the secretary of the Army, at the annual Association of the U.S. Army conference in Washington.

To move faster on cyber, Kwon said several initiatives must be met. First on the list is artificial intelligence-driven mission command, with the top goal to achieve decision dominance.

Using AI will be critical to operate at the speed needed to react to cyber threats, Kwon said.

Next is pushing commanders to see a data-centric Army, treating data as a strategic asset. Kwon likened data to the new ammunition, both critical pieces of the larger warfighting puzzle.

Third is that training must evolve beyond linear scenarios and move into a contested environment, and leaders must push units to outpace the enemy in their battlefield actions.

Lastly, Kwon noted that with the division as the new unit of action in the Army’s plans for large-scale combat operations, much experimenting will happen at that level. As a division goes into the dirt at one of the combat training centers, cyber work will feature prominently in what leaders will learn from each rotation.

Pugh warned that commanders can’t see cyber as an isolated capability performed by certain cyber commands or units. Division and corps commanders have access to these tools and must use them in conjunction with other assets, such as electronic warfare, he said.

Rey tied together the non-kinetic nature of the fight, noting that units must be able to fight from their forts to their ports, all of which will require resilient cyber capabilities and practices.

The plane landed safely on Sunday, but Bulgarian authorities said they suspected Russia was behind the interference.

“It is taken very seriously,” said NATO Secretary-General Mark Rutte during a news conference in Luxembourg with the duchy’s prime minister and defense minister. “I can assure you that we are working day and night to counter this, to prevent it, and to make sure that they will not do it again.” He did not elaborate.

Neither Russia nor von der Leyen has commented publicly on the incident. The EU and NATO are separate entities with different sets of member countries, but Europe’s security is a vital issue for both.

Rutte said the jamming was part of a complex campaign by Russia of “hybrid threats” like cutting of undersea cables in the Baltic Sea, a plot to assassinate a German industrialist, and a cyberattack on the National Heath Service in the United Kingdom.

“I have always hated the words hybrid because it sounds so cuddly, but hybrid is exactly this jamming of commercial airplanes, with potentially disastrous effects,” he said.

The Associated Press has plotted almost 80 incidents on a map tracking a campaign of disruption across Europe blamed on Russia, which the head of Britain’s foreign intelligence service has described as “staggeringly reckless.” Since Moscow’s full-scale invasion of Ukraine in 2022, Western officials have accused Russia and its proxies of staging dozens of attacks and other incidents, ranging from vandalism to arson and attempted assassination.

The interference from Russia includes jamming and spoofing. Jamming means a strong radio signal overwhelms communications, whereas spoofing misleads a receiver into thinking it is in a different location or in a past or future time period.

“The threat from the Russians is increasing every day. Let’s not be naive about it: this might also involve one day Luxembourg, it might come to the Netherlands,” Rutte said. “With the latest Russian missile technology for example, the difference now between Lithuania on the front line and Luxembourg, The Hague or Madrid is five to 10 minutes. That’s the time it takes this missile to reach these parts of Europe.”

The whole continent was under “direct threat from the Russians,” he warned. “We are all on the eastern flank now, whether you live in London or Tallinn.”

Bulgaria will not investigate the jamming of von der Leyen’s plane because “such things happen every day,” Bulgarian Prime Minister Rosen Zhelyazkov said Tuesday.

He said it was one of the side effects of Russia’s war in Ukraine and had occurred across Europe.

Associated Press writer Veselin Toshkov in Sofia, Bulgaria contributed to this report.

Air Force Weather started its digital transformation in 2017 amid a broader U.S. government push to migrate away from siloed data centers to more secure, efficient and capable could-based environments. As the Air Force’s largest special-purpose data processing node — crunching around 80 terabytes of data, or the equivalent of 6.6 billion pages of text, a day — the organization was a “big fat target” for cyber threats, according to Fred Fahlbusch, Air Force Weather’s data domain officer and chief of the weather resources, programs, data and cybersecurity division.

So, with fresh funding in hand, it began the process of migrating its operations from physical servers at the 557th Weather Wing at Offutt Air Force Base in Nebraska to its new digital home in the cloud. It contracted with Amazon for cloud services and shifted its high-performance computing capability from Offutt to Oak Ridge National Laboratory in Tennessee, which could better accommodate its power and cooling needs.

The transition has been difficult on multiple levels, but Fahlbusch and other officials told Defense News the organization is on the verge of completing its change to the cloud by next year — a milestone they say is likely to spike demand for the operational weather products the organization provides to its user base.

“If we deliver what people want, then the demand signals are going to explode,” Fahlbusch said.

That demand, while a sign of success, presents a whole new set of challenges for Air Force Weather. More utilization requires more resources, Fahlbusch said, and the organization will likely have to take a new approach to funding the services it provides.

Virtual Private Cloud

According to Rod Grady, chief architect and cloud integrator at the Air Force Life Cycle Management Center’s Weather Systems Branch, it’s not data that makes Air Force Weather indispensable. In fact, while the organization draws data from Air Force and Space Force radars and satellites, much of the weather information it relies on comes from unclassified sources like the National Oceanic and Atmospheric Administration, or NOAA, and commercial firms.

The organization’s value proposition, instead, is in what it does with that data.

“You can go out and buy weather data from many different sources,” Grady said. “But what makes us unique is the application of that data to support mission decision-making. And we have services and capability on enclaves that are being exploited every day by the millions.”

The organization, which is the go-to source for environmental data tools for the Air Force, Space Force, Army and intelligence community, sees around 30 million requests a month for weather products from users operating in classified enclaves.

One such tool gives military aircraft access to severe weather data when operating in areas without radar coverage. A hydro-intelligence product developed by Air Force Weather provides detailed flood mapping and forecasting. And a web-based application uses a network of terminals to pull data from U.S. and international weather satellites and make it available to users.

Moving operations to the cloud will help the organization develop more tools at a faster rate and increase its processing abilities, Grady said. It also will make its products available to more users.

The office is already seeing the initial impact of cloud operations on its products, as its transition to what it calls the Virtual Private Cloud, or VPC, has been incremental. One illustration of this, Grady said, is a capability called the Operational Data Layer, which allows a forecaster to create mission-specific tools out of large data sets.

The office started experimenting with the capability in 2010, but the technology proved too challenging and the project lost momentum. In 2020, amid the transition to VPC, Air Force Weather revived the effort with a new plan to use cloud computing processes to develop the data tool. This time, it went from initial concept to operations in less than six months and the organization is now expanding the capability for more applications.

“That product is now the number one hit product and is integrated into a number of applications across the Air Force and the DOD on the unclassified level and higher enclaves,” Grady said.

Fahlbusch said Air Force Weather’s products and delivery timelines will only improve once it’s fully leveraging VPC. And a better product combined with a cloud-based environment that allows users outside its typical base to access its tools will likely mean a significant increase in utilization.

“Once we’re in a cloud environment and available to people we’re not currently available to, there could be an explosion of demand from end users we’re not even aware of,” Fahlbusch said.

Managing demand

The office welcomes the larger user base but is also preparing for the challenges that will come with managing more demand. Fahlbusch said funding is his biggest concern.

The organization, whose fiscal 2025 funding sits around $250 million — not including personnel costs – could request more money to operate in the cloud, but that’s a difficult ask when the defense budget is tightening, Fahlbusch said.

It could also shift to a fee-for-service model, which puts the cost burden on the offices using its services. While the products the organization provides differentiate it from other sources, the cost could prove to be a barrier for users who are also seeing their budgets shrink, he said.

Fahlbusch is not alone in his concerns. Managing the cost and distribution of cloud services has been a challenge for Pentagon agencies that have embraced cloud operations over the last decade without aligning their processes under an enterprise-level management approach. In an effort to address these issues, the Defense Department released a Cloud Financial Operations Strategy last year designed to help both providers and users better navigate the cloud.

Air Force Weather is applying that strategy as it prepares for demand, according to Grady and Fahlbusch — monitoring usage more tightly and putting mechanisms in place to make sure it’s not sending duplicative information or larger amounts of data than necessary.

In practice, that means working more closely with users to get a better understanding of how much data they actually need for a particular task, which is often less than they originally requested, Fahlbusch said.

“We can work with them, and we can tailor it,” he said. “I think that will help defray some of the costs smartly and help curate what the end user really needs — targeted curation, if you will, versus asking for everything. That’ll kill us.”

Workforce, funding cuts

Air Force Weather’s concerns about managing the supply and demand that come with providing a more effective and capable product are set against the backdrop of the Trump administration’s deep cuts to the federal civilian workforce. The Department of Government Efficiency has also targeted contracting expertise and agencies like NOAA — both of which the organization relies on to provide high-quality weather products.

The office has seen some indirect impacts of those reductions, with civilian workforce cuts hitting the Air Force Life Cycle Management Center, which provides acquisition and program management support. Fahlbusch said cuts to contractors and federally funded research and development centers — which provide crucial expertise to support Air Force Weather’s cloud transition — could also impact the agency.

“We need these specialists not as consultants but working the front lines as ‘digital grease monkeys,’ doing the coding, data curation, integration and other digital functions to help optimize weather integration into weapon systems,” a spokesman added in a June 4 statement following the interview.

Deep staffing and budget reductions at NOAA could also impact Air Force Weather, Fahlbusch said, giving it fewer reliable data sources from which to draw.

It’s not yet clear what the scope of that impact might be. Fahlbusch noted that while funding and personnel reductions could make it harder for Air Force Weather to continue to support military readiness and planning, the team understands the push for fiscal responsibility and will do what it can to realign manpower and funding if required.

“If there are things we have to stop doing to focus on those warfighting priorities, we will do it . . . consistent with the president and [secretary of defense] guidance,” he said. “Because that’s what we’re all about.”

The finalized deal comes as Leidos is implementing a new “NorthStar” strategy that identifies cyber as one of five primary growth areas for the company.

“Kudu’s ability to generate new cyber capabilities with AI perfectly complements our strategy to rapidly grow differentiated offensive cyber technology capabilities,” Leidos CEO Tom Bell said in a statement. “This acquisition underlines Leidos’s commitment to continue to build smarter full-spectrum cyber capabilities, so that the U.S. and its allies dominate the cyber warfighting domain.”

Kudu has around 170 employees and is based in Chantilly, Virginia. The firm has customers in the Defense Department and the intelligence community.

Roy Stevens, president of Leidos’ national security sector, told Defense News the firm was particularly drawn to Kudu’s expertise in identifying cyber vulnerabilities, developing code that can exploit flaws in a system and then create an electronic-warfare or other effect.

He specifically highlighted the firm’s application of AI to exploit an adversary system.

“We really believe warfare is moving more and more in that direction,” Stevens said. “That’s a place where they have a unique position.”

Those strengths, he said, directly align with the opportunities and gaps Leidos has identified within its own cyber portfolio, and the company expects the acquisition will help it get after its growth targets 18 months faster than it may have otherwise.

Leidos today has a large defensive and offensive cyber footprint and provides support to DOD and intelligence customers, including the National Security Agency. The company in April won a $390 million contract to provide signals intelligence to the NSA, and last June the Air Force and Space Force re-upped their enterprise IT contract with a $738 million award to Leidos.

The decision allows Switzerland to become part of the Estonian-led Cyber Ranges Federations project under the EU’s Permanent Structured Cooperation (PESCO) framework, marking a notable advance in Swiss–EU military cooperation. This comes despite Bern’s famously longstanding policy of strict military neutrality.

Switzerland had applied to join the project in October of last year, shortly after submitting an application for another joint project focused on military mobility. Two formalities remain before becoming a full project member: Estonia must invite Switzerland to the cooperation, and Bern needs a so-called administrative arrangement with the EU governing formalities such a data exchange and other parameters.

The Swiss government welcomed this week’s EU decision, saying that the country “will take part in the European PESCO project.”

Switzerland has beefed up its own cyber defense capabilities in recent years with its Swiss Cyber Training Range and a Cyber-Defence Campus.

The EU’s Cyber Ranges Federations initiative seeks to centralize capacity, pool unique services and automate processes across member states, reducing manual workload during exercises and accelerating the development of advanced cybersecurity technologies.

Austria, Belgium, Bulgaria, Finland, France, Italy and Luxembourg are already members of the project, in addition to Estonia.

Under PESCO’s third-state participation rules established in 2020, non-EU countries may join individual projects if they share EU values and pose no threat to member states’ security interests. The Council confirmed that Switzerland meets the required political, legal and substantive criteria and will bring “substantial added value and mutual benefit” to the federation, it said in a press release.

The Council retains oversight of third-state involvement and may adjust conditions should security considerations evolve, ensuring alignment with the EU’s collective defense objectives.

Swiss defense planners have balanced these new engagements with Bern’s policy of armed neutrality, with federal officials calling cooperation in PESCO initiatives “ad hoc collaboration on specific projects which are thematically in the interest of both parties and which do not create critical dependencies for neutrality.”

Participation in the cyber project enables Switzerland to contribute – and benefit from – expertise and infrastructure without entangling the country too deeply in broader EU defense commitments, from Bern’s point of view. The Swiss government said that “participation will take place selectively and on a needs-oriented basis.”

The latest project represents part of Switzerland’s broader strategic approach to selective participation in PESCO projects that align with its defense interests while maintaining neutrality.

It’s not Switzerland’s first brush with EU defense initiatives. In January, the government received the green light to join an EU-led military mobility project, which it applied for in September 2024. The Military Mobility project aims to simplify and standardize national cross-border military transport procedures, enabling swift movement of military personnel and assets throughout the EU via road, rail, sea, or air. Other non-EU countries, such as the UK, Northway, the USA and Canada are also part of this project.

In addition to deepening engagement with the EU, Switzerland has also been a member of NATO’s partnership for peace since the 1990s, as has its neutral eastern neighbor, Austria.

Hardline neutrality defenders have long taken issue with Swiss engagement on military projects beyond its own borders. Their criticism received new urgency in the aftermath of Russia’s attack on Ukraine in 2022. Last year, a civil movement garnered more 130,000 certified signatures to organize a national referendum on strengthening Switzerland’s international neutrality. The referendum organizers specifically want to prevent what they see as a gradual erosion of Switzerland’s traditional neutrality through strengthened international defense cooperation.

The situation arrives as concerns grow that lax regulations are allowing AI programmers to dodge responsibility for an algorithmic actor’s perpetration of emotional abuse or “no-marks” cybertorture. Notably, a teenager allegedly died by suicide — and several others endured mental distress — after conversing with self-learning voicebot and chatbot “companions” that dispensed antagonizing language.

Now, seven years after writing about physical torture in “The Rise of A.I. Interrogation in the Dawn of Autonomous Robots and the Need for an Additional Protocol to the U.N. Convention Against Torture,” privacy attorney Amanda McAllister Novak sees an even greater need for bans and criminal repercussions.

In the wake of the use of “chatbots, voicebots, and other AI-enabled technologies for psychologically-abusive purposes — we have to think about if it’s worth having AI-questioning systems in the military context,” said Novak, who also serves on the World Without Genocide board of directors. Novak spoke to Defense News on her own behalf.

Royal Reff, a spokesperson for the U.S. Defense Counterintelligence and Security Agency (DCSA), which screens individuals holding or applying for security clearances, said in a statement that the agency’s AI voice agent was “expressly developed” to “mitigat[e] gender and cultural biases” that human interviewers may harbor.

Also, “because systems like the interviewer software” are “currently being used worldwide, including within the U.S. by state and local law enforcement, we believe that it is important for us to understand the capabilities of such systems,” he stated.

DCSA spokesperson Cindy McGovern added that the development of this technology continues because of its “potential for enhancing U.S. national security, regardless of other foreign applications.”

After hearing the rationale for the project, Novak registered concerns about the global race for dominance in military AI capabilities.

Odds are high that “no matter how much the government cares about proper training, oversight and guardrails,” cybercriminal organizations or foreign-sponsored cybercriminals may hack into and weaponize military-owned AI to psychologically hurt officers or civilians, she said.

“Online criminal spaces are thriving” and “harming the most vulnerable people,” Novak added.

Investors are betting $500 billion that data centers for running AI applications will ultimately secure world leadership in AI and cost savings across the public and private sectors. The $13 billion conversational AI market alone will nearly quadruple to $50 billion by 2030, as the voice generator industry soars from $3 billion to an expected $40 billion by 2032. Meanwhile, the U.S. Central Intelligence Agency has been toying with AI interrogators since at least the early 1980s.

DCSA officials publicly wrote in 2022 that whether a security interview can be fully automated remains an open question. Preliminary results from mock questioning sessions “are encouraging,” officials noted, underscoring benefits such as “longer, more naturalistic types of interview formats.”

Reff stated that the agency “would not employ any such system that was not first sufficiently studied to understand its potential,” including “negative consequences.”

Researchers let volunteers raise concerns after each test, and none have indicated that the voicebot interviewer harmed them psychologically or otherwise, he added.

However, experiments with a recent prototype barred volunteers with mental disorders from participating, whereas actual security interviews typically do not exempt individuals with mental health conditions from questioning, noted Renée Cummings, a data science professor and criminal psychologist at the University of Virginia.

Cummings said that research and testing can never account for every psychological variable that may influence the emotional or cognitive state of the interviewee.

In laboratory or field-testing scenarios, “you can’t expose someone to extreme torture by an avatar or by a bot to see the result,” said Cummings, who also sits on the advisory council of the AI & Equality Initiative at the Carnegie Council for Ethics in International Affairs. “We need to focus on gaining a more sophisticated understanding about the impacts of algorithms on the psychological state, our emotions and our brain — because of AI’s ability to impact in real-time.”

Screening the voicebot screeners

An early 1980s CIA transcript of a purported conversation between an AI program and an alleged spy highlights the national security world’s long-held interest in automating human-source intelligence, or HUMINT, collection.

Developers designed the chatbot’s algorithms to analyze correlations between suspected spook “Joe Hardesty’s” behaviors and words or phrases in Hardesty’s responses to open-ended questions.

From this analysis, the chatbot pegged some of the suspect’s vulnerabilities, including topics that may hit a raw nerve:

AI: What do you tend to think is my interest in your government?

Hardesty: I think you would do in my government if you had a chance.

AI: You are prone to think I would do in your government because?

Hardesty: You are anti-democratic.

In 1983, the CIA quipped that Joe Hardesty is fortunate that “should the probing get too discomforting, he will have an option that will not be available to him in a true overseas interview situation — he can stop the questions with a flick of the ‘off’ switch.”

Today, Bradley Moss, a lawyer who represents whistleblowers accused of leaking, is not laughing at DCSA’s vision of fully-automated interviews.

“There would have to be someone observing” the avatar “that can stop the process, that can come into the room and insert a human element to it,” he cautioned.

Then again, other critics note that a human may not want to intervene. A situation with “a human overseer monitoring the chatbot is fine, unless the human also believes that verbal torture is a good thing. And that is a real possibility,” said Herb Lin, a senior research scholar for cyber policy and security at Stanford University who once held several security clearances.

Despite a global ban on torture and evidence that rapport-building interviews are more efficient than ones gunning for confessions, intelligence officials continue subjecting interviewees to inhumane treatment.

Novak, the data privacy attorney and anti-genocide activist, said the idea that excessive force gets the job done may encourage some government contractors to set performance metrics for voice algorithms that permit bullying.

“While AI has a tremendous ability to bring out the best in human potential, it also has the ability to exacerbate the worst of human potential and to elevate that hostility,” she added.

Parents contend that voicebot and chatbot companions, trained on content from online forums, show just that.

Character AI, the firm that offers virtual relationships with bots patterned after icons and influencers, profits from users glued to their phones. Each user, on average, talks two hours a day with various virtual friends. Character AI officials have stated that, after learning of concerns, the company reduced the likelihood of users encountering suggestive content and now warns them of hour-long conversations.

Novak commented that the accusations against Character AI illustrate the ability of self-learning voicebots to amplify indecency in training data and to distance trainers from accountability for that offensiveness.

In the national security realm, developers of military AI voice systems that may potentially abuse or terrorize operate in a virtually lawless no-man’s land. AI ethics policies in the world’s most powerful country offer nearly no guidance on de-escalating abusive conversations.

Early U.S. responsible and ethical AI policies in the national security space focused on accuracy, privacy and minimizing racial and ethnic biases — without regard for psychological impact. That policy gap widened after President Donald Trump revoked many regulations in an executive order aimed at accelerating private sector innovation.

In general, psychological suffering is often “brushed away” as a mere allegation since “it leaves no obvious physical scars.”

The United Nations initially did not address “psychological torture” under the 1984 Convention Against Torture. Not until five years ago did the term take on significance.

A 2020 UN report now defines “torture,” in part, as any technique intended or designed to purposefully inflict severe mental pain or suffering “without using the conduit” of “severe physical pain or suffering.”

At bottom, psychological torture targets basic human needs. The interrogation techniques arouse fear; breach privacy or sexual integrity to kindle shame or suicidal ideation; alter sound to heighten disorientation and exhaustion; foster and then betray sympathy to isolate; and instill helplessness.

When an algorithmic audiobot is the conduit of cruelty, neither the UN treaty nor other international laws are equipped to hold humans accountable, Novak said. Without binding rules on the use of AI in the military context, “bad actors will deny any such torture was ‘intentional’” — a required element for criminal responsibility that the law created before AI matured, she explained.

Looking to the future, Nils Melzer, author of the 2020 UN report and a former UN special rapporteur on torture, warned that the interpretation of psychological torture must evolve in sync with emerging technologies, “such as artificial intelligence,” because cyber environments provide “virtually guaranteed anonymity and almost complete impunity” to offenders.