{"id":1665818,"date":"2026-02-20T23:25:00","date_gmt":"2026-02-21T04:25:00","guid":{"rendered":"https:\/\/bugaluu.com\/news\/?p=1665818"},"modified":"2026-02-20T23:25:00","modified_gmt":"2026-02-21T04:25:00","slug":"as-pentagon-races-to-deploy-ai-operational-challenges-highlight-risks-3","status":"publish","type":"post","link":"https:\/\/bugaluu.com\/news\/as-pentagon-races-to-deploy-ai-operational-challenges-highlight-risks-3\/1665818\/","title":{"rendered":"As Pentagon Races to Deploy AI, Operational Challenges Highlight Risks"},"content":{"rendered":"

As Pentagon Races to Deploy AI, Operational Challenges Highlight Risks<\/span><\/p>\n

\n

Authored by Autumn Spredemann via The Epoch Times<\/a> (emphasis ours),<\/em><\/p>\n

Artificial intelligence (AI) is often framed as a force multiplier that can accelerate decision-making and produce valuable information. <\/strong>Meanwhile, AI deployment exercises have yielded mixed results, highlighting challenges such as systems stalling and unpredictable software outside controlled environments.<\/p>\n

A U.S. soldier holds a drone in the Pentagon parking lot in Arlington, Va., on June 14, 2025. Samuel Corum\/Getty Images<\/em><\/a><\/p>\n

Some defense insiders believe that AI tools also introduce new safety and escalation risks if not developed, evaluated, and trained correctly.<\/strong><\/p>\n

Over the past year, U.S. military testing has demonstrated that some AI systems are failing in the field. In May 2025, Anduril Industries worked with the U.S. Navy on the launch of 30 AI drone boats, all of which ended up stuck idling in the water after the systems rejected their inputs.<\/p>\n

A similar setback occurred in August 2025 during the company\u2019s test of its Anvil counterdrone system. The resultant mechanical failure caused a 22-acre fire in Oregon, according to a Wall Street Journal report.<\/p>\n

Anduril responded to the reported AI test failures, calling them \u201ca small handful of alleged setbacks at government experimentation, testing, and integration events.\u201d<\/p>\n

\u201cModern defense technology emerges through relentless testing, rapid iteration, and disciplined risk-taking,<\/strong>\u201d Anduril stated on its website. \u201cSystems break. Software crashes. Hardware fails under stress. Finding these failures in controlled environments is the entire point.\u201d<\/p>\n

But some say the challenges AI faces in the national security landscape should not be taken lightly. Problems such as brittle AI models and building on the wrong kind of training data can create systems that do not perform as expected in a battlefield scenario.<\/p>\n

\u201cThis is why military-grade AI, purpose-built for national security use cases and the warfighter, is critical<\/strong>,\u201d Tyler Saltsman, founder of EdgeRunner AI, told The Epoch Times.<\/p>\n

Saltsman\u2019s company has active research and development contracts with the U.S. military. He said AI systems are not typically designed for warfighting.<\/p>\n

\u201c[AI models] may choose to refuse or deflect certain questions or tasks if those requests do not comply with the AI system\u2019s own rules<\/strong>,\u201d Saltsman said. \u201cA model refusing to provide guidance to a soldier in combat or giving biased responses rather than operationally relevant responses can have life-or-death implications.\u201d<\/p>\n

Scenarios such as the one Saltsman described can start with the wrong kind of training data.<\/p>\n

A U.S. Army staff sergeant operates an Anduril Ghost X unmanned aircraft system during Exercise Balikatan 25 in Itbayat, Philippines, on April 22, 2025. While artificial intelligence is often framed as a force multiplier, deployment exercises have produced mixed results, including system stalls and unpredictable software performance outside controlled environments. Pfc. Peter Bannister\/U.S. Army<\/em><\/a><\/p>\n

Data Dilemma<\/h2>\n

Jeff Stollman, who has worked with defense contractors as an independent consultant and is familiar with a range of products and services used by the military and intelligence communities, said much of \u201cthe data needed has not been collected historically.\u201d<\/p>\n

\u201cAnd because internet data is typically of limited value and internet-based models can\u2019t be run on isolated classified networks, military and intelligence users will need to collect their own new data<\/strong>,\u201d Stollman told The Epoch Times.<\/p>\n

He said there are three categories of training data used by the defense and armed forces communities, all of which have different hurdles.<\/p>\n

Offering an example of a sustainment\u2014or maintenance\u2014data challenge, Stollman said that collecting this type of information typically requires adding sensors that can record the data needed to predict malfunctions and failures.<\/strong><\/p>\n

\u201cThis includes measuring temperature, vibration, friction, the amount of wear on various parts,\u201d he said. \u201cThis is an expensive undertaking. Sensors aren\u2019t free. They add weight and volume to space and weight-constrained platforms such as aircraft and spacecraft.\u201d<\/p>\n

This type of data collection is offloaded to a database because of limited onboard computer resources. Although that sounds logical at first, the problem is the time it can take.<\/p>\n

\u201cFor platforms like ships and submarines, windows for transmission of such data, which might give away the position of the platform, are limited,\u201d Stollman said. \u201cAs a result, data may not be accessible for months at a time.\u201d<\/p>\n

A drone of an AI-based drone system is pictured during a presentation in Eberswalde, Germany, on March 27, 2025. Ralf Hirschberger\/AFP via Getty Images<\/em><\/a><\/p>\n

Another challenge of AI integration is reliability<\/strong>. Issues such as AI \u201challucinations\u201d and poor decisions can be amplified in adversarial environments.<\/p>\n

\u201cThe most dangerous assumption is that AI can distinguish between legitimate inputs and adversarial manipulation,\u201d Christopher Trocola, founder of ARC Defense Systems, told The Epoch Times.<\/p>\n

He cited the July 2025 experiment in which AI-powered, cloud-based platform Replit\u2019s \u201cvibe coding\u201d ended with an AI assistant panicking and trying to cover its tracks. The AI coding assistant reportedly deleted a live production database, fabricated thousands of fake records, and created misleading status messages.<\/p>\n

\u201cMilitary applications amplify these vulnerabilities catastrophically<\/strong>,\u201d Trocola said.<\/p>\n

He explained that three critical AI assumptions can fail under adversarial pressure: prompt injection resistance, hallucination control, and intent recognition.<\/p>\n

This is when adversaries can manipulate AI through carefully crafted inputs designed to override instructions, generate false information, or indicate that malicious inputs are benign.<\/p>\n

\u201cThis represents what\u2019s known as distribution shift: AI trained in controlled environments failing catastrophically when deployed in real-world adversarial contexts<\/strong>,\u201d Trocola said.<\/p>\n

Saltsman said this highlights the importance of building AI models with military applications in mind.<\/p>\n

\u201cMost commercial AI systems are black boxes,\u201d he said. \u201cWe don\u2019t know what data trained the models. We don\u2019t know what guardrails or biases were baked into the models. And we don\u2019t know if our data is truly secure. All of this is highly problematic in national security settings.\u201d<\/p>\n

Risk Evaluation<\/h2>\n

Stollman noted that generative AI\u2014which is already used in U.S. intelligence and defense\u2014is \u201cplagued\u201d with problems such as hallucinations. However, it is also the most practical kind of AI for military operations.<\/p>\n

\u201cGenerative AI is useful in areas such as reconnaissance, where it is necessary to identify installations and activities from data collected by various sensors: photos, radar, sonar, etc.,\u201d Stollman said. \u201cIt can also be used to support decision-making.\u201d<\/p>\n

A consultant instructs the Advanced Artificial Intelligence Command Course at Marine Corps Base Camp Lejeune in North Carolina on Dec. 12, 2025. Lance Cpl. Payton Walley\/U.S. Marine Corps<\/em><\/a><\/p>\n

\u201cFor example, drones or missiles could be given autonomy of action to overcome signal jamming that prevents their being controlled remotely by humans,\u201d he said. \u201cBut before such autonomy can be deployed, it is necessary to anticipate all the failure modes that could lead to undesirable consequences.\u201d<\/p>\n

Saltsman said he agrees that AI development and deployment must be carefully balanced with long-term risk evaluation.<\/p>\n

\u201cBut make no mistake, we are in an AI war against China, and we must win the race<\/strong>,\u201d he said.<\/p>\n

He noted that if China\u2019s AI models and hardware dominate the market, the United States could become dependent on the Asian nation for critical technologies.<\/p>\n

\u201cTherefore, it is a national security imperative that we accelerate the pace of AI development while also balancing the risks,\u201d Saltsman said.<\/p>\n

In 2025, the United Nations said that the use of AI in warfighting was no longer a hypothetical future scenario. The U.N. also stressed the risks and consequences of AI system failures in this capacity.<\/p>\n

\u201cWithout rigorous safeguards, it risks undermining international humanitarian law,\u201d the agency stated<\/a>.<\/p>\n

\u201cComplex battlefields already test human judgment in distinguishing between combatants and civilians; for machines, the challenge is even greater, particularly in urban settings where civilians and fighters often intermingle.\u201d<\/p>\n

Xpeng\u2019s next-gen Iron humanoid robot speaks to media during a showroom tour at its headquarters in Guangzhou, Guangdong Province, China, on Nov. 5, 2025. Tyler Saltsman said that if China\u2019s AI models and hardware dominate the market, the United States could become dependent on the Asian nation for critical technologies. Jade Gao\/AFP via Getty Images<\/em><\/a><\/p>\n

Trocola said he shares concerns that AI deployment in the military and defense sectors is outpacing risk assessment.<\/p>\n

\u201cDocumented patterns suggest this creates systematic vulnerabilities,\u201d he said. \u201cIndustry data shows [70 percent to 80 percent] of AI projects fail due to organizational readiness gaps.\u201d<\/p>\n

The Department of War AI Acceleration Strategy launched<\/a> in January, which emphasizes rapid deployment to counter strategic competitors.<\/p>\n

Read the rest\u00a0here<\/a>…<\/strong><\/em><\/p>\n<\/div>\n

Tyler Durden<\/a><\/span>
\nFri, 02\/20\/2026 – 18:25<\/span><\/p>\n

\u200bhttps:\/\/www.zerohedge.com\/military\/pentagon-races-deploy-ai-operational-challenges-highlight-risks<\/a>\u00a0<\/p>\n","protected":false},"excerpt":{"rendered":"

As Pentagon Races to Deploy AI, Operational Challenges Highlight Risks Authored by Autumn Spredemann via The Epoch Times (emphasis ours), Artificial intelligence (AI) is often…<\/p>\n","protected":false},"author":0,"featured_media":1665816,"comment_status":"closed","ping_status":"closed","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[1],"tags":[],"class_list":["post-1665818","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-news","wpcat-1-id"],"_links":{"self":[{"href":"https:\/\/bugaluu.com\/news\/wp-json\/wp\/v2\/posts\/1665818","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/bugaluu.com\/news\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/bugaluu.com\/news\/wp-json\/wp\/v2\/types\/post"}],"replies":[{"embeddable":true,"href":"https:\/\/bugaluu.com\/news\/wp-json\/wp\/v2\/comments?post=1665818"}],"version-history":[{"count":0,"href":"https:\/\/bugaluu.com\/news\/wp-json\/wp\/v2\/posts\/1665818\/revisions"}],"wp:attachment":[{"href":"https:\/\/bugaluu.com\/news\/wp-json\/wp\/v2\/media?parent=1665818"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/bugaluu.com\/news\/wp-json\/wp\/v2\/categories?post=1665818"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/bugaluu.com\/news\/wp-json\/wp\/v2\/tags?post=1665818"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}