This funding call is aimed at developing next-generation human-AI dialogue systems specifically tailored for defence applications. The call addresses the fast-paced evolution of generative artificial intelligence, large language models and chatbots, and their potential for high-impact use in defence domains such as intelligence, strategic planning, tactical operations, and life-cycle support. However, with the benefits of these advanced technologies come significant challenges – notably, the risk of generating false or misleading information, known as hallucinations, and the necessity to adapt the systems to the stringent and unique requirements of defence settings.

Applicants are invited to develop multilingual dialogue systems that not only demonstrate state-of-the-art performance but also integrate advanced features such as explainable AI, robust handling of sensitive or classified information, and the capability to continuously learn from user feedback without requiring constant developer intervention. In addition to ensuring that the systems can interact both in writing and by voice, proposals must also cover all EU official languages, thereby fostering an inclusive approach that supports a wide range of military users across EU Member States and Associated Countries.

The call emphasizes the importance of research on dialogue systems as a fundamental step towards generating new knowledge and underpinning disruptive technologies in the defence sector. To ensure that these systems meet operational demands, each consortium funded under this call will be required to integrate their developed technological modules into demonstrators with user-friendly interfaces. These demonstrators are expected to run locally in near real-time with no perceptible lag, even when dealing with classified information. As part of the evaluation process, all proposals must participate in testing campaigns organized by a third-party consortium. These campaigns will include experimental performance measurements, debriefing workshops, and exchanges with other stakeholders regarding the evaluation plans.

Beyond the technical specifications, the call seeks to enhance trust and the broader acceptance of generative AI within defence systems by ensuring data sovereignty and security. It requires that the systems justify their answers through explainability, handle classified data through access-controlled learning, and continuously evolve based on direct user supervision. The successful demonstration of these capabilities will contribute to improved decision-making in complex operational scenarios, ultimately strengthening the European Defence Technological and Industrial Base (EDTIB) and advancing technological autonomy in the defence arena.

Furthermore, the proposals must clearly outline work packages, tasks, and deliverables so that the evaluation of progress is straightforward and transparent. The integration of various technological modules into realistic demonstrators is a key aspect of the project, as these demonstrators not only serve as proof of concept but also generate realistic data that can be used for further system development. By bridging the gap between cutting-edge AI research and its practical application in defence, this call is set to play a crucial role in ensuring that future systems are both highly performant and trusted by users.

Opening: 18.02.2025
Deadline(s): 16.10.2025