French Air Force plans loyal wingman flight in 2028 in sovereign AI push

PARIS — France’s Air Force, together with the country’s Directorate General for Armament and the agency for defense artificial intelligence, plans to fly crewed fighter jets in collaboration with combat drones in 2028, part of a project to experiment with AI in combat aviation dubbed Hypairion.

The project seeks to develop open, modular architecture for combat aircraft and sovereign AI technology that gives the government greater control over the tech, rather than having to rely on companies for critical systems, said Col. Cédric, head of the Air Force capability-development office, in a press briefing here on Thursday.

As AI becomes more powerful, the likes of Airbus and BAE Systems in Europe and General Atomics and Anduril in the United States are leveraging the technology to develop collaborative combat aircraft, also referred to as “loyal wingmen.” While many of those projects are company-driven, the French project would keep some of the critical technology under state control.

“What Hypairion is really about is equipping the Air Force and our platforms with architectures that are open, modular and owned by the government,” Col. Cédric said. He was identified only by his first name, in line with the Armed Forces Ministry’s policy for military staff. “Our goal is to move away from proprietary systems that lock down sensors and weaponry.”

The Ministerial Agency for Defence AI will be in charge of providing sovereign AI technology for the project in several areas, particularly command and control as well as mission autonomy, said Col. Thibault, deputy director for AI applications at the agency, known as Amiad. He described mission autonomy as AI that assists pilots with mission planning, in-flight decisions and post-mission analysis.

Amiad already uses AI to coordinate different platforms in the French Army’s Pendragon robotics project, and proposes to create an industrial hub for Hypairion where both traditional aerospace firms and smaller startups can interact with the project, an approach tested in Pendragon.

“The point of the Hypairion project is for the French state to retain control of the overall architecture and to be able to integrate the most capable startups or industrial companies,” Col. Thibault said. He said companies might join and leave the project, while Hypairion retains its capabilities in-house.

The Air Force is modifying two Mirage 2000s fighters at Mont-de-Marsan airbase to become AI test beds, adding a new computer that will allow to test “all kinds of use cases” starting by the end of the year, according to Col. Cédric. Mont-de-Marsan in southwest France is home to the Escadron de chasse et d’expérimentation, a dedicated fighter squadron for testing and evaluation.

Development will be incremental, and the first stage of Hypairion won’t necessarily be about having an operational capability, but will look at areas such as energy consumption and system interfacing, Col. Cédric said. First use cases being considered include using AI to optimize flight paths or tactical decisions such as when to deploy air-to-air or air-to-ground weapons.

The unmanned aircraft set to fly in 2028 will be a test platform and may not be the future wing-man drone planned for France’s Rafale jet as part of the fighter’s future F5 standard, though the experimentation will help pick a suitable platform later in the decade, Col. Cédric said.

The DGA has set a target data of 2033 for the first version of the Rafale F5, but will work on collaborative platforms within the framework of the current F4 standard before then, said General Engineer of Armaments Sébastien, in charge of the Rafale program at the directorate.

The goal of integrating AI on the Rafale is to help crews manage the “absolutely phenomenal cognitive load” to maintain future operational superiority, he said. The armaments engineer cited the Talios laser-designation pod from Thales, set to be delivered to the armed forces in 2027, as a concrete application of AI, able to scan an area and provide a fighter crew a list of potential targets.

Col. Cédric described a hypothetical future in which a fighter crew would carry out a mission in the morning, in between missions identify an adaptation against an enemy countermeasure, and take off again in the afternoon with updated software. That will require system architectures to evolve, he said.

The French Air Force officer likened future manned and unmanned platforms to “flying computers” on which the keyboard, screen or software could be changed without compromising security. The goal of the architecture being developed within Hypairion is to separate the flight-safety aspect from the ability to carry out missions, he said.

The Air Force doesn’t envisage a future with only unmanned aircraft, for both ethical and operational reasons, according to the colonel, who said the decision to deploy weapons remains a human one, and AI can make mistakes. “Humans can be augmented or supplemented by AI, but the assessment of the operational situation, tactical judgment, and so on, these are still entirely the domain of humans.”

As the Air Force prepares the Rafale F5 standard, “a true mid-life overhaul of the aircraft,” focus areas include adequacy of hardware resources, as the system needs to be able to handle complex calculations, Sébastien at the DGA said. Another focus is system openness, meaning “the armed forces must be able to tailor artificial intelligence according to their own needs and so dispose of an open system.”

The Rafale’s F4 standard already includes an initial level of openness through a tablet connected to the mission system, and the F5 standard will seek to broaden the scope of data available to the tablet, give it access to the mission system’s computing resources, and improve the ability to interface with onboard communication systems and data links, according to the DGA engineer.

“This in particular will make it possible, and without waiting for the deployment of the F5 standard, to control autonomous combat platforms from the Rafale,” Sébastien said.

While France relies on Dassault Aviation and other major industrial partners for the Rafale to develop AI in the fighter’s major systems, the armed forces want to ensure operational autonomy, meaning having “the ability to tailor the AI to the specific operational needs of the theater,” the engineer said.

Hypairion will work on a road map covering usage concepts, doctrine, scaling up the work force as well as the actual equipment, resulting in a first full-scale trial of collaborative combat drones flying alongside fighter jets in 2028, according to Col. Cédric.

“The challenge for the Air and Space Force is ultimately to determine what a fighter squadron will look like in 2035 or 2040, once all these systems are potentially integrated into the squadron,” Col. Cédric said. “Today, it’s actually quite rare to come across computer scientists in a fighter squadron. Perhaps tomorrow, that will be a daily occurrence.”

Rudy Ruitenberg is a Europe correspondent for Defense News. He started his career at Bloomberg News and has experience reporting on technology, commodity markets and politics.

巴黎——法国空军与该国的武器总局和国防人工智能机构计划在 2028 年将有人驾驶战斗机与作战无人机协同飞行，这是名为 Hypairion 的战斗航空人工智能试验项目的一部分。

空军能力发展办公室主任塞德里克上校周四在此举行的新闻发布会上表示，该项目旨在为战斗机开发开放式模块化架构和自主人工智能技术，使政府能够更好地控制这项技术，而不是依赖公司提供关键系统。

随着人工智能技术的日益强大，欧洲的空客和BAE系统公司，以及美国的通用原子公司和安杜里尔公司等企业正在利用这项技术开发协同作战飞机，也被称为“忠诚僚机”。虽然许多此类项目都是由企业主导的，但法国的这个项目将使一些关键技术保持在国家控制之下。

“Hypairion项目的真正意义在于为空军及其平台配备开放、模块化且由政府拥有的架构，”塞德里克上校说道。根据武装部队部对军事人员的政策，他只被提及了名字。“我们的目标是摆脱那些封闭传感器和武器系统的专有系统。”

国防部人工智能机构（简称Amiad）人工智能应用副主任蒂博上校表示，该机构将负责为该项目提供自主研发的人工智能技术，涵盖多个领域，特别是指挥控制和任务自主性。他将任务自主性定义为人工智能辅助飞行员进行任务规划、飞行中决策和任务后分析。

Amiad 已经利用人工智能在法国陆军的 Pendragon 机器人项目中协调不同的平台，并提议为 Hypairion 创建一个工业中心，让传统的航空航天公司和较小的初创企业都能参与到该项目中来，这种方法已经在 Pendragon 中进行了测试。

“Hypairion项目的意义在于，法国政府能够掌控整体架构，并整合最优秀的初创企业或工业公司，”蒂博上校说道。他表示，企业可以随时加入或退出该项目，而Hypairion将始终保持其内部能力。

据塞德里克上校透露，美国空军正在蒙德马桑空军基地对两架幻影2000战斗机进行改装，使其成为人工智能测试平台。改装后将加装一台新计算机，用于测试“各种应用场景”，预计将于年底前完成。蒙德马桑位于法国西南部，是法国战斗机测试与评估中队（Escadron de chasse et d'expérimentation）的所在地，该中队专门负责测试和评估工作。

塞德里克上校表示，Hypairion的研发将循序渐进，第一阶段的目标未必是具备作战能力，而是着眼于能源消耗和系统接口等领域。目前正在考虑的首批应用案例包括利用人工智能优化飞行路线或战术决策，例如何时部署空对空或空对地武器。

塞德里克上校表示，计划于 2028 年试飞的无人机将是一个测试平台，可能并非法国“阵风”战斗机未来 F5 标准中计划使用的僚机无人机，但此次试验将有助于在本十年晚些时候选择一个合适的平台。

DGA已将Rafale F5首款机型的目标交付日期定为2033年，但在此之前，DGA将在现有F4标准的框架内开展合作平台研发工作。负责Rafale项目的DGA武器总工程师Sébastien表示。

他表示，在“阵风”战斗机上集成人工智能的目标是帮助机组人员应对“极其巨大的认知负荷”，从而保持未来的作战优势。这位武器工程师以泰雷兹公司生产的Talios激光指示吊舱为例，指出人工智能的具体应用。该吊舱将于2027年交付给美军，能够扫描目标区域并为战斗机机组人员提供潜在目标列表。

塞德里克上校描述了一个设想的未来场景：战斗机机组人员上午执行任务，在任务间隙找出应对敌方反制措施的改进方案，下午带着更新后的软件再次起飞。他表示，这将需要系统架构的演进。

这位法国空军军官将未来的有人和无人平台比作“飞行计算机”，其键盘、屏幕或软件都可以在不影响安全性的前提下进行更换。他表示，Hypairion 正在开发的架构的目标是将飞行安全与任务执行能力分离。

这位上校表示，出于伦理和作战方面的考虑，空军并不设想未来只使用无人机作战。他指出，武器部署的决定仍然由人做出，人工智能也会犯错。“人工智能可以增强或补充人类的能力，但对作战形势的评估、战术判断等等，仍然完全属于人类的范畴。”

空军正在筹备“阵风”F5标准升级，这相当于“飞机的一次真正中期大修”。法国国防采购局（DGA）的塞巴斯蒂安表示，重点领域包括硬件资源的充足性，因为该系统需要能够处理复杂的计算。另一个重点是系统开放性，这意味着“武装部队必须能够根据自身需求定制人工智能，从而拥有一个开放的系统”。

据DGA工程师称，阵风战斗机的F4标准已经通过连接到任务系统的平板电脑实现了初步的开放性，而F5标准将寻求扩大平板电脑可获取的数据范围，使其能够访问任务系统的计算资源，并提高与机载通信系统和数据链路的接口能力。

“这将使‘阵风’战斗机无需部署 F5 标准即可控制自主作战平台，”塞巴斯蒂安说。

这位工程师表示，虽然法国依靠达索航空公司和其他主要工业合作伙伴为“阵风”战斗机的主要系统开发人工智能，但武装部队希望确保作战自主性，这意味着“能够根据战区的具体作战需求定制人工智能”。

据塞德里克上校称，Hypairion 将制定一份路线图，涵盖使用概念、理论、扩大劳动力规模以及实际设备，最终在 2028 年进行首次大规模的协同作战无人机与战斗机并肩飞行试验。

“空军和太空军面临的最终挑战是，一旦所有这些系统都可能整合到战斗机中队中，那么到2035年或2040年，战斗机中队将会是什么样子，”塞德里克上校说。“如今，在战斗机中队里遇到计算机科学家其实相当罕见。也许明天，这就会成为日常现象。”

鲁迪·鲁伊滕贝格是《防务新闻》的欧洲记者。他的职业生涯始于彭博新闻社，拥有报道科技、大宗商品市场和政治方面的经验。