Citation and evidence

Defense AI

22 min full readUpdated 30 references

This article's verification

Report a problem with this article

Edit article
More

Use this article

Raw MarkdownExplore connections

Improve this page

Revision historyDiscuss this article

Topics

AI EthicsAI Policy & RegulationDefense AI

Cite this article

Defense AI (also military AI) is the use of artificial intelligence for military and national security purposes, including intelligence analysis, logistics, cyber operations, command and planning, uncrewed vehicles and weapons. The term also covers the government offices that buy and manage these systems, the companies that build them, and the policies and international talks that set limits on their use. The Congressional Research Service (CRS) has described the field as one in which the United States and other nations are developing AI for "a range of military functions," with research underway in intelligence collection and analysis, logistics, cyber operations, information operations, command and control, and semiautonomous and autonomous vehicles [1].

Much public debate about defense AI concerns weapons that could select and engage targets on their own, a subject covered in detail in autonomous weapons. Many fielded programs are analysis and decision-support tools rather than weapons; Project Maven, for example, was built to help analysts sort drone video [1]. Since 2025 the field has also come to include generative AI and large language models bought directly from commercial frontier labs, which led to a 2026 legal dispute between the US Department of Defense and Anthropic over the limits the company placed on military use of its models [3][7].

Applications

CRS groups military uses of AI into several areas [1]. The table summarizes them with examples drawn from public programs.

AreaWhat AI is used forPublic examples
Intelligence, surveillance and reconnaissanceSorting imagery and video, flagging objects and activity for analystsProject Maven and the Maven Smart System [1][12]
Logistics and maintenancePredicting equipment failures, managing supplyJAIC's predictive maintenance initiative [1][4]
Cyber and information operationsNetwork defense and attack; detecting manipulated mediaDARPA media forensics work [1]
Command, control and planningFusing data across services, drafting and wargaming courses of actionJoint All-Domain Command and Control; DIU's Thunderforge [1][13]
Uncrewed and autonomous vehiclesNavigation, swarming, terminal guidanceSea Hunter, Navy swarm-boat tests, Replicator [1][5]
Air combatAI agents flying aircraftDARPA Air Combat Evolution (ACE) [25][26]
Enterprise and office workGenerative AI assistants for staffGenAI.mil [8]

Expanded article table

CRS expected intelligence to be a particularly useful area for AI because of the large data sets available for analysis. It described Project Maven as an effort to put computer vision algorithms into intelligence cells that would comb through footage from uncrewed aerial vehicles and automatically identify hostile activity for targeting, so that analysts would spend less time sifting through drone video [1]. In logistics, the Joint Artificial Intelligence Center (JAIC) chose predictive maintenance and humanitarian assistance and disaster relief as its first two National Mission Initiatives in 2019 [1]. For information operations, CRS noted both the threat of deepfakes and DARPA's Media Forensics (MediFor) project, which developed initial tools for identifying AI-produced forgeries [1].

CRS reported that DARPA completed testing of its Anti-Submarine Warfare Continuous Trail Unmanned Vessel prototype, Sea Hunter, in early 2018 before transitioning the program to the Navy, and that in November 2016 the Navy tested an AI-enabled swarm of five uncrewed boats that cooperatively patrolled a 4-by-4-mile section of the Chesapeake Bay and intercepted an "intruder" vessel [1]. In air combat, the US Air Force Test Pilot School and DARPA announced in April 2024 that AI agents flying the X-62A VISTA test aircraft had fought dogfights against a human pilot in 2023 under the ACE program; the teams made more than 100,000 lines of flight-critical software changes across 21 test flights, and safety pilots on board, who could disengage the AI, never had to do so during the dogfights [25]. Shield AI lists the X-62 among the aircraft that run its Hivemind autonomy software [26].

"Inside the Pentagon's AI Revolution", video from Bloomberg Television, published October 19, 2025. Watch on YouTube.

United States programs

Early strategy and the Third Offset

Autonomous systems were a central focus of the Obama administration's "Third Offset Strategy," a framework for preserving the US military's technological edge over competitors [1]. The January 2018 National Defense Strategy named artificial intelligence as one of the technologies that would "ensure [the United States] will be able to fight and win the wars of the future," and the Department of Defense (DoD) also produced a classified AI strategy [1]. CRS reported that DoD's unclassified AI investment grew from just over $600 million in FY2016 to $2.5 billion in FY2021, including autonomy, and that the department reportedly maintained more than 600 active AI projects; DARPA separately announced a $2 billion multiyear investment in more than 20 AI programs [1].

Project Maven

Project Maven was launched in 2017 to develop AI that can detect, tag and track objects or people of interest in images and video from surveillance aircraft, satellites and other sources [12]. CRS described it in 2020 as a "spearhead initiative" that had used AI algorithms to identify insurgent targets in Iraq and Syria [1]. In 2022 responsibility for Maven was split among the National Geospatial-Intelligence Agency, the Chief Digital and AI Office and the Office of the Under Secretary of Defense for Intelligence and Security [12].

Maven was also the first large public dispute between a technology company and the Pentagon. In May 2018 Google employees successfully lobbied the company to withdraw from Project Maven and avoid further collaboration with DoD [1]. In June 2018 chief executive Sundar Pichai published AI principles that listed, among applications Google would not pursue, "weapons or other technologies whose principal purpose or implementation is to cause or directly facilitate injury to people," while saying the company would continue to work with governments and the military in other areas [17]. Google updated those principles in February 2025; the announcement, written by James Manyika and Demis Hassabis, set out three core tenets without the earlier list of applications the company would not pursue and said that "democracies should lead in AI development" [18].

The software built around Maven grew into a large contract for Palantir. In May 2024 the Army awarded Palantir a five-year contract for the Maven Smart System initially worth $480 million, and in May 2025 the Pentagon raised the ceiling by $795 million to nearly $1.3 billion through 2029 [12]. The director of the National Geospatial-Intelligence Agency said at the time that Maven had more than 20,000 active users, and DefenseScoop reported that Palantir had signed a deal with NATO for a version of the system [12].

From JAIC to the Chief Digital and AI Office

DoD set up the Joint Artificial Intelligence Center in a June 27, 2018, memo, placing it under the Chief Information Officer; section 238 of the FY2019 National Defense Authorization Act also provided for the center, to "develop, mature, and transition artificial intelligence technologies into operational use" [4]. In May 2020 JAIC awarded an $800 million contract to Booz Allen Hamilton to "bring AI to the battlefield" [1]. On December 8, 2021, Deputy Secretary of Defense Kathleen Hicks directed the creation of the Office of the Chief Digital and Artificial Intelligence Officer (CDAO) as JAIC's successor, reporting directly to the Deputy Secretary, with the Defense Digital Service reporting through it and the Chief Data Officer operationally aligned to it [4].

On August 14, 2025, Deputy Secretary Steve Feinberg moved CDAO under the Under Secretary for Research and Engineering, a change CRS called "an effective demotion" [4]. Retired Lieutenant General Jack Shanahan, JAIC's first director, criticized the move, saying that when an office that reported to the deputy secretary is moved elsewhere, "the message to the force is loud and clear: This isn't a priority," while supporters described it as the next step in an "AI-first push" [4]. On November 13, 2026, the department renamed the office the Super Intelligence Office (SIO), saying the change did not alter its mission and aligned the name with President Trump's push to replace references to artificial intelligence with "super intelligence" across the federal government; the office, led by the department's top AI official Cameron Stanley, oversees programs including the Maven Smart System [28].

Replicator

Deputy Secretary Hicks unveiled Replicator on August 28, 2023, with the goal of fielding thousands of uncrewed systems by August 2025 [5]. Its first line of effort aimed at "all-domain, attritable autonomous" systems, meaning relatively cheap systems whose loss the military can tolerate, and a second line, announced in September 2024, focused on countering small drones [5]. DoD officials said the initiative drew on lessons from Ukraine and was meant to offset the Chinese military's "advantage in mass" [5]. Publicly confirmed selections included AeroVironment's Switchblade 600 loitering munition and Anduril's Altius-600 and Ghost-X drones [5]. DoD requested $500 million for Replicator in FY2025, and CRS reported in January 2026 that, according to one former defense official, DoD had fielded only hundreds of systems by the summer 2025 deadline [5].

Frontier AI and the AI Acceleration Strategy

In 2025 the department contracted directly with frontier model developers. In July 2025 CDAO announced awards to Anthropic, Google and xAI, each with a ceiling of $200 million, following an equal award to OpenAI in June; the stated aim was to "develop agentic AI workflows across a variety of mission areas" [7]. Breaking Defense noted that the combined $800 million was a maximum that depended on CDAO exercising every option [7]. Anthropic had already released Claude Gov in June 2025, models "built exclusively for U.S. national security customers" and designed to refuse less when handling classified material [16].

On September 5, 2025, Executive Order 14347 allowed the department to use "Department of War" as a secondary name in non-statutory contexts; the statutory name remains Department of Defense until Congress changes it [10]. In January 2026 Secretary Pete Hegseth issued a memorandum setting out an AI strategy meant to make the department an "AI-first warfighting force" [8]. The strategy rests on three pillars (warfighting, intelligence and enterprise operations) and seven "Pace-Setting Projects": Swarm Forge, Agent Network, Ender's Foundry, Open Arsenal, Project Grant, GenAI.mil and Enterprise Agents [8][9]. GenAI.mil is the department's platform for staff access to commercial models, including Gemini and Grok [8].

Language models are also being applied to military planning. In March 2025 the Defense Innovation Unit gave Scale AI a prototype contract for Thunderforge, which uses large language models, AI simulations and agent-based wargaming to help planners generate and test courses of action, with Anduril contributing its Lattice software platform, Microsoft providing the language models, and first deployments planned at Indo-Pacific Command and European Command [13].

Policy framework

DoD AI Ethical Principles

DoD adopted five AI ethical principles on February 24, 2020, after recommendations from the Defense Innovation Board that followed 15 months of consultation [6]. The principles state that AI should be responsible, equitable, traceable, reliable and governable, and they apply to both combat and non-combat functions [6]. Directive 3000.09 later tied the use of AI in autonomous and semi-autonomous weapons to these principles [2].

Directive 3000.09

DoD Directive 3000.09, "Autonomy in Weapon Systems," first issued in November 2012 and updated in January 2023, sets US policy on autonomous weapons [2]. It defines an autonomous weapon as one that, "once activated, can select and engage targets without further intervention by a human operator," and requires that all such systems allow commanders and operators to exercise "appropriate levels of human judgment over the use of force" [2]. CRS stresses that, contrary to frequent news reports, US policy does not prohibit the development or use of lethal autonomous weapons; instead it requires a senior-level review before formal development and again before fielding, which the Deputy Secretary can waive for "urgent military need" [2]. Congress has added notification rules: the FY2024 NDAA requires notice of any change to the directive within 30 days, the FY2025 NDAA requires an annual report on approval and deployment of lethal autonomous weapons through 2029, and the FY2026 NDAA requires notice of any waiver [2].

Industry

Suppliers range from established defense contractors to newer defense technology firms and general-purpose AI labs. CRS noted in 2020 that most AI development happens in the commercial sector, which forces the defense acquisition system to adapt and means many commercial products need significant changes before they are useful to the military [1].

CompanyDefense AI work (examples)
PalantirMaven Smart System for DoD and NATO [12]
AndurilAltius-600 and Ghost-X drones selected for Replicator; Lattice software; counter-drone partnership with OpenAI [5][13][15]
Shield AIHivemind autonomy software for aircraft including the X-62 and V-BAT [26]
Scale AIThunderforge planning prototype for DIU [13]
Anthropic, Google, OpenAI, xAICDAO frontier AI awards of up to $200 million each [7]

Expanded article table

Other AI Wiki articles cover additional defense technology companies, including Helsing, Skydio, Saronic and Vannevar Labs.

Frontier labs' published policies on military use changed during the same period. The Intercept reported in January 2024 that OpenAI had removed a ban on "military and warfare" uses from its usage policy, while keeping a prohibition on using its services to develop or use weapons [14]. In December 2024 Anduril and OpenAI announced a partnership on counter-unmanned aircraft systems, to explore how AI models trained on Anduril's data could "rapidly synthesize time-sensitive data, reduce the burden on human operators, and improve situational awareness" [15].

The Anthropic dispute

According to CRS, reports indicated that Claude was used in the January 2026 operation to capture Venezuelan President Nicolás Maduro and that Anthropic then asked about the department's use of the model; the Pentagon reportedly responded by asking Anthropic and other AI companies to allow use of their models for "all lawful purposes" [3]. Anthropic was reportedly willing to adapt its usage policies but would not permit two uses, mass domestic surveillance and fully autonomous weapon systems; chief executive Dario Amodei said such weapons "may prove critical for our national defense. But today, frontier AI systems are simply not reliable enough to power fully autonomous weapons" [3]. CRS also notes that DoD is not publicly known to use Claude or any other frontier model inside an autonomous weapon [3].

On February 27, 2026, President Donald Trump directed federal agencies to stop using Anthropic's technology, and Hegseth directed the department to designate the company a supply-chain risk and bar contractors from working with it [3]. Anthropic sued on March 9 [3]. The two courts, applying two different statutes, then reached different results. On August 27 the Northern District of California held that the designation under 10 U.S.C. 3252 violated the statute, was arbitrary and capricious, and was unlawful retaliation under the First Amendment, and enjoined it [3]. On September 25 a 2-1 majority of a D.C. Circuit panel upheld a separate exclusion order under the Federal Acquisition Supply Chain Security Act (41 U.S.C. 4713), finding "ample support" for the department's conclusion that keeping Claude in its systems posed a national security risk [3][11]. As of CRS's October 5, 2026 update, DoD could continue removing Anthropic products from its systems unless the full D.C. Circuit or the Supreme Court ruled otherwise, and Anthropic said it was "considering all options, including further review" [3]. Anthropic then petitioned for rehearing en banc before all active judges of the D.C. Circuit [30].

Other countries and alliances

China. CRS reported that China released a plan on July 20, 2017, to lead the world in AI by 2030, created a Military-Civil Fusion Development Commission the same year to speed the transfer of AI from commercial firms and research institutions to the military, and was focused mainly on faster, better-informed decision-making and autonomous military vehicles [1]. China's 2019 defense white paper said that "intelligent warfare is on the horizon" and called for the armed forces to "speed up the development of intelligent military" [20]. See AI in China for the civilian side.

Russia. Less than two months after China's 2017 plan, Vladimir Putin said that "whoever becomes the leader in this field will rule the world" [1]. CRS described Russian military AI research as focused heavily on robotics and semiautonomous and autonomous vehicles [1].

NATO. NATO published an AI strategy on October 22, 2021, with six principles of responsible use: lawfulness; responsibility and accountability; explainability and traceability; reliability; governability; and bias mitigation [19].

Ukraine. A March 2025 CSIS study by Kateryna Bondar found that autonomy "is not yet present on the battlefield in the war in Ukraine," but that platforms with basic autonomous functions such as navigation or targeting were in use, that AI significantly enhanced specific functions such as drone footage analysis and target recognition, and that autonomous navigation could "raise the target engagement success rate from around 10 to 20 percent to around 70 to 80 percent" [21]. The study reported that Ukraine produced about 2 million drones in 2024 [21]. US officials have cited this use of large numbers of cheap drones as a model for Replicator [5].

Use in armed conflict

Besides Maven's use against insurgents in Iraq and Syria [1] and drone autonomy in Ukraine [21], detailed reporting concerns Israel's war in Gaza and the 2026 US campaign against Iran. In April 2024 +972 Magazine and Local Call published an investigation by Yuval Abraham, based on six Israeli intelligence officers, reporting that an AI system called Lavender had marked as many as 37,000 Palestinians as suspected militants and potential targets, that, according to one source, officers would typically spend about 20 seconds on each target before authorizing a strike, and that the system was known to err in about 10 percent of cases [22]. The IDF Spokesperson denied the existence of such a kill list and denied using AI to incriminate targets, describing its systems as "auxiliary tools that assist officers in the process of incrimination" and saying an independent examination by an intelligence analyst is required to verify each target [22].

In Operation Epic Fury, the US air and missile campaign against Iran that began on February 28, 2026, senior US officials said the Pentagon relied on the Maven Smart System to identify top-priority targets and help choose the weapons used against them [29]. The Pentagon's chief digital and AI officer, Cameron Stanley, said in March 2026 that Maven fused radar, satellite and drone imagery, communications and other data into a single picture and generated courses of action, so that commanders could go from identifying a target to striking it "all from one system" [29]. The Arms Control Association reported that Anthropic's Claude had been merged with the Maven technology in late 2024 to help generate and rank target lists, and that critics blamed overreliance on AI for targeting errors and civilian casualties; whether AI played a role in the February 28 strike on a girls' elementary school in Minab, which killed more than 170 people, was unclear, and The New York Times reported officials as saying that error was unlikely to have resulted from new technology [29].

International governance

Discussion of military AI at the United Nations has centered on lethal autonomous weapons in the Convention on Certain Conventional Weapons, where talks began in 2014 and became a formal Group of Governmental Experts in 2017 [2]. The United States does not support a ban, while CRS counts about 30 countries and 165 non-governmental organizations that have called for one [2]. These talks and the related General Assembly resolutions are covered in autonomous weapons.

Wider military AI, not only weapons, is the subject of two other tracks. The US State Department issued the Political Declaration on Responsible Military Use of Artificial Intelligence and Autonomy on November 9, 2023, which 58 states had endorsed as of November 27, 2024 [2]. The Summits on Responsible AI in the Military Domain (REAIM) were held in The Hague in 2023, Seoul in 2024 and A Coruña, Spain, on February 4-5, 2026 [23][24]. The 2023 summit ended with a non-binding Call to Action agreed by 57 states, including the United States and China [27]. Reuters reported that the non-binding action plan from the 2024 Seoul summit was backed by about 60 countries including the United States but not China, and that at the 2026 summit only 35 of the 85 attending countries signed the joint declaration, with neither the United States nor China among them [23].

Debate

Supporters of military AI argue that it can make operations faster and more precise. The US government has argued that automated target identification, tracking, selection and engagement can strike military objectives more accurately and with less risk of collateral damage [2], and Defense Secretary Mark Esper, announcing the department's AI ethical principles, said the United States and its allies must accelerate AI adoption to "prevail on future battlefields" while keeping its commitment to "responsible and lawful behavior" [6].

Critics raise concerns about human control, accountability and escalation, which are set out in autonomous weapons. Some senior officers have shared part of that caution: then-Vice Chairman of the Joint Chiefs General Paul Selva told the Senate Armed Services Committee in 2017, "I do not think it is reasonable for us to put robots in charge of whether or not we take a human life," while adding that the military would still need to study the technology because adversaries were pursuing it [1]. A separate line of criticism concerns execution rather than ethics: analyst Bill Greenwalt told the House Armed Services Committee in 2023 that the Pentagon's acquisition system was "simply not capable of acting on the proposed timelines contemplated in the Replicator program," and CRS has noted the gap between Replicator's goal of thousands of systems and the hundreds reportedly fielded [5]. The Anthropic case raised the question of who sets limits on a government's use of a commercially supplied model; CRS notes that some lawmakers have called for Congress to set rules for DoD's use of AI and autonomous weapons [3].

Continue reading

More to read

Topics:AI EthicsAI Policy & RegulationDefense AI

References

  1. ^1 ^2 ^3 ^4 ^5 ^6 ^7 ^8 ^9 ^10 ^11 ^12 ^13 ^14 ^15 ^16 ^17 ^18 ^19 ^20 ^21 ^22 ^23 ^24Kelley M. Sayler, Congressional Research Service. "Artificial Intelligence and National Security". CRS Report R45178, updated November 10, 2020. everycrsreport.com/...R45178
  2. ^1 ^2 ^3 ^4 ^5 ^6 ^7 ^8 ^9Congressional Research Service. "Defense Primer: U.S. Policy on Lethal Autonomous Weapon Systems". CRS In Focus IF11150, updated March 26, 2026. everycrsreport.com/...IF11150
  3. ^1 ^2 ^3 ^4 ^5 ^6 ^7 ^8 ^9 ^10Kelley M. Sayler, Congressional Research Service. "Pentagon-Anthropic Dispute over Autonomous Weapon Systems: Potential Issues for Congress". CRS Insight IN12669, updated October 5, 2026. everycrsreport.com/...IN12669
  4. ^1 ^2 ^3 ^4 ^5Kelley M. Sayler, Congressional Research Service. "Realignment of DOD's Chief Digital and AI Officer (CDAO)". CRS Insight IN12615, updated December 1, 2025. everycrsreport.com/...04a484c4afcd2d00daeb376d.pdf
  5. ^1 ^2 ^3 ^4 ^5 ^6 ^7 ^8 ^9Congressional Research Service. "DOD Replicator Initiative: Background and Issues for Congress". CRS In Focus IF12611, updated January 21, 2026. everycrsreport.com/...IF12611
  6. ^1 ^2 ^3U.S. Department of Defense. "DOD Adopts Ethical Principles for Artificial Intelligence". Press release, February 24, 2020. defense.gov/...nciples-for-artificial-intelligence
  7. ^1 ^2 ^3 ^4Sydney J. Freedberg Jr. "Anthropic, Google and xAI win $200M each from Pentagon AI chief for 'agentic AI'". Breaking Defense, July 14, 2025. breakingdefense.com/...gon-ai-chief-for-agentic-ai
  8. ^1 ^2 ^3 ^4Elodie Collins. "Pete Hegseth Introduces War Department Strategy to Accelerate AI Adoption". ExecutiveGov, January 13, 2026 (updated January 21, 2026). executivegov.com/...dow-ai-adoption-strategy-hegseth
  9. ^Office of the Under Secretary of War for Research and Engineering. "Release of the War Department's AI Acceleration Strategy". CTO.mil, undated. cto.mil
  10. ^The White House. "Restoring the United States Department of War" (Executive Order 14347). September 5, 2025. whitehouse.gov/...-united-states-department-of-war
  11. ^Amy L. Fuentes, David S. Black, Ann E. McRitchie, Tanner N. Slaughter and Marissa C. Serafino. "D.C. Circuit Upholds Exclusion of Anthropic from DOW Supply Chain Under FASCSA". Holland & Knight, September 29, 2026. hklaw.com/...on-of-anthropic-from-dow-supply-chain
  12. ^1 ^2 ^3 ^4 ^5 ^6Brandi Vincent. "'Growing demand' sparks DOD to raise Palantir's Maven contract to more than $1B". DefenseScoop, May 23, 2025. defensescoop.com/...smart-system-contract-increase
  13. ^1 ^2 ^3 ^4Defense Innovation Unit. "DIU's Thunderforge Project to Integrate Commercial AI-Powered Decision-Making for Operational and Theater-Level Planning". March 5, 2025. diu.mil/...e-commercial-ai-powered-decision-making
  14. ^Sam Biddle. "OpenAI Quietly Deletes Ban on Using ChatGPT for 'Military and Warfare'". The Intercept, January 12, 2024. theintercept.com/...open-ai-military-ban-chatgpt
  15. ^1 ^2Anduril Industries. "Anduril Partners with OpenAI to Advance U.S. Artificial Intelligence Leadership and Protect U.S. and Allied Forces". December 4, 2024. anduril.com/...lligence-leadership-and-protect-u-s
  16. ^Anthropic. "Claude Gov models for U.S. national security customers". June 6, 2025. anthropic.com/...r-u-s-national-security-customers
  17. ^Sundar Pichai. "AI at Google: our principles". Google blog, June 7, 2018. blog.google/...ai-principles
  18. ^James Manyika and Demis Hassabis. "Responsible AI: Our 2024 report and ongoing work". Google blog, February 4, 2025. blog.google/...onsible-ai-2024-report-ongoing-work
  19. ^NATO. "Summary of the NATO Artificial Intelligence Strategy". October 22, 2021. nato.int/...official_texts_187617
  20. ^State Council Information Office of the People's Republic of China. "China's National Defense in the New Era". July 24, 2019. english.www.gov.cn/...t_WS5d3941ddc6d08408f502283d
  21. ^1 ^2 ^3Kateryna Bondar. "Ukraine's Future Vision and Current Capabilities for Waging AI-Enabled Autonomous Warfare". Center for Strategic and International Studies, March 6, 2025. csis.org/...s-waging-ai-enabled-autonomous-warfare
  22. ^1 ^2Yuval Abraham. "'Lavender': The AI machine directing Israel's bombing spree in Gaza". +972 Magazine and Local Call, April 3, 2024. 972mag.com/lavender-ai-israeli-army-gaza
  23. ^1 ^2Victoria Waldersee (Reuters). "AI軍事利用の国際サミット、中国と米国は共同宣言に署名せず" (International summit on military AI: China and US do not sign joint declaration). Newsweek Japan, February 6, 2026. newsweekjapan.jp/...312567
  24. ^Spanish Ministry of Foreign Affairs, European Union and Cooperation. "Spanish MFA closes world's largest debate on responsible use of military AI in A Coruña". Press release, February 5, 2026. exteriores.gob.es/...-responsable-de-la-IA-militar
  25. ^1 ^2Chase Kohler, 412th Test Wing. "USAF Test Pilot School and DARPA announce breakthrough in aerospace machine learning". DVIDS, April 17, 2024. dvidshub.net/...468721
  26. ^1 ^2 ^3Shield AI. "Hivemind". Accessed November 2026. shield.ai/hivemind
  27. ^Tobias Vestner and Juliette François-Blouin. "Globalizing Responsible AI in the Military Domain by the REAIM Summit". Just Security, March 13, 2023. justsecurity.org/...ary-domain-by-the-reaim-summit
  28. ^Mikayla Easley. "Pentagon rebrands principal AI shop as 'Super Intelligence Office'". Breaking Defense, November 13, 2026. breakingdefense.com/...s-super-intelligence-office
  29. ^1 ^2 ^3Michael Klare. "AI Plays Major Role in the War on Iran". Arms Control Today, Arms Control Association, May 2026. armscontrol.org/...ai-plays-major-role-war-iran
  30. ^Kevin Frazier. "Why Anthropic Is Asking for a Rehearing En Banc". Lawfare, October 15, 2026. lawfaremedia.org/...asking-for-a-rehearing-en-banc

Improve this article

Add missing citations, update stale details, or write a clearer explanation. Anyone can edit: changes go live when you publish, and every edit is checked by an editor.

2 revisions · v2 · 4,332 words · full history

Fact-checks are independent of edits: a reviewer re-verifies the article against its sources and stamps the date. How we verify

Research and drafting on this wiki are AI-assisted, under named human editorial standards. How AI is used here

Reviewer note: Independent fact-check before publication on November 18, 2026: 95 claims checked against 34 cited and independent sources; 3 material and 7 minor corrections applied before publication.

Cite this page: AI Wiki. "Defense AI." aiwiki.ai, updated 18 Nov 2026, fact-checked 18 Nov 2026. CC BY 4.0. https://aiwiki.ai/wiki/defense_ai

Edit article