Poland and France ramp up AI investments for military edge

The armed forces, as an organisation, are constantly evolving. On one hand, they are inherently conservative; on the other, since the 19th century, generals have been seeking ways to gain a technological edge. Sometimes they succeed more effectively, sometimes less so. Occasionally, pioneers gain at least temporary dominance on the battlefield, while at other times, premature innovations advantage the adversary. Currently, one of the most promising technologies is artificial intelligence (AI).

During the AI Action Summit organised in Paris on 10-11 February, numerous interesting declarations were made. Among others, there was an announcement to reduce legal regulations to facilitate the development of AI in Europe (at the EU level as well).

France alone, according to its president Emmanuel Macron, intends to invest €109 (£91) billion in AI over the next few years. This is an impressive amount, somewhat larger (in terms of GDP conversion) than the nearly $500 (£395) billion President Donald Trump announced for the same purpose a few weeks earlier. The French plan to use these funds to design and build data centres supporting AI development.

The French Armed Forces will also benefit from this revolution. Bertrand Rondepierre, Director of AI (AMIAD) at the Ministry of Armed Forces, stated in an interview with "Politico" that France must undertake comprehensive actions towards formulating AI algorithms and models and their training. Their own artificial intelligence algorithms will be essential for maintaining military sovereignty in the near future.

The AMIAD agency, established in May 2024, has an annual budget of £251 million and employs over 100 people. The work is broad in scope, ranging from AI for drones (and counter-drone measures) to language models for document processing and supporting military planners. In September 2025, the agency will receive its own supercomputer, the most powerful in Europe, though it will be supplied by Hewlett-Packard in collaboration with Orange. Rondepierre highlighted that the actual monopoly of NVIDIA on essential hardware for ensuring high computational power significantly complicates AI development because it is dependent on Washington's policies. However, Paris plans to increase its range of sovereignty, albeit mainly in the software domain.

Of course, France is not currently a leader in the development of artificial intelligence, as it does not produce its own chips. The USA and China are particularly more advanced, but significantly more countries are working on their solutions: Germany, India, Israel, South Korea... These efforts have a strictly defined goal; they are not done just for art's sake. So why do soldiers need AI?

Artificial intelligence in the military world has a range of applications at many levels. During the Paris summit, Rondepierre pointed out that AI is supposed to facilitate understanding of the battlefield situation, including operational or logistical dimensions. AI is responsible for analysing data, classifying it according to significance, and forwarding it to officers responsible for using it. This way, generals will have access to all the data at once, but AI will ensure that the most important data reaches them first, allowing for faster processing and use in decision-making.

One could say that AI will serve as a gatekeeper, deciding the order of messages. In the future, this function will likely expand, possibly towards evaluating military plans in the context of available data. AI might also find use in cryptography or cybersecurity.

On a lower level, AI is already available to soldiers. It is most renowned as an indispensable element of the drone world. AI allows for effective target searching (e.g., by extracting target silhouettes from the background based on information from electro-optical heads), planning flight paths, etc. It also offers a chance to break through some electronic warfare measures that jam drone-command centre communication. In such cases, AI provides the possibility to continue the mission or at least return to base.

In the future, more advanced versions of these algorithms could practically pilot unmanned combat aircraft supporting manned machines - they will be the "brains" of the so-called unmanned wingmen. This is not the only application of AI at the tactical level.

Even today, it's difficult to imagine underwater warfare in the navy without AI. In a series about the Orka programme, I wrote that modern submarines are extremely difficult to detect. Being very quiet, they blend into background noise. Artificial intelligence allows the extraction of the correct sound - generated, for instance, by the engine room - from that noise, enabling detection of the large "predator".

In ground forces, AI is expected to assist crews in choosing optimal driving routes, managing the vehicle (again: information classification), automating certain processes, etc. For example, the Merkava Mk IV Barak was initially intended to communicate automatically with other similar vehicles so that tanks could mutually protect each other in case of a threat without crew involvement. The tank's AI was supposed to communicate with the crew in a female voice, which during tests was found to be preferable to a male voice, as crews ignored it less frequently. However, it is unclear how many solutions were ultimately implemented in the latest Israeli Defence Forces tank.

Poland allocates small amounts for the development of AI, and this is no different in terms of military applications. It should be noted that Deputy Prime Minister and Minister of Digital Affairs Krzysztof Gawkowski mentioned that artificial intelligence should not be a subject of an arms race.

Despite such evident challenges, some progress is occasionally made. Among the practical outcomes, drones are again the most apparent example, such as systems of the Warmate family or FlyEye produced by the WB Group. Thanks to experience gained from their use in Ukraine, Polish company programmers can continuously improve them, including enhancements in AI limiting jamming effects or in drone and loitering munition guidance. The capabilities of the ZSSW-30 unmanned turret remain unclear - according to some reports, it also has AI algorithms supporting crews in combat.

The announcement of the establishment of the Centre for Artificial Intelligence Implementation offers some hope for more significant outcomes, as announced by Major General Karol Molenda, commander of Cyberspace Defence Forces. The new formation will be responsible for the development and implementation of AI solutions for military needs. According to reports, the Centre will work on solutions following the aforementioned trends. Polish "AI in uniform" would thus support cyber defence, commanders in the planning process, and command, etc. The Ministry of National Defence's "AI Departmental Strategy until 2039" published in October 2024 defines the necessity of implementing AI into the Polish Armed Forces as the highest priority. Is it truly such a priority? This will be revealed by future budgets and the amount of funding allocated for this purpose. For now, reasons for optimism remain rather limited.