What happened

Ukraine is beginning cooperation with the international company Beyond PL — a well-known Nvidia partner in infrastructure for artificial intelligence. According to the Ministry of Digital Transformation, within this work specialized small language models (SLM) will be developed, adapted for specific state sectors and deployment in a secure environment.

Beyond PL will provide computational resources for training these models, and will also help form the state AI Factory — infrastructure with reliable data centers for storing and processing data.

"Cooperation with Beyond PL will allow state models to be trained in a secure environment and lay the infrastructure for the sovereign use of AI in public services."

— Ministry of Digital Transformation

Why it matters

This is not just a technical project — it is a question of national infrastructure and security. Proprietary models give the state control over data, reduce dependence on external clouds, and make it possible to establish rules for validation, audit, and the protection of personal data.

For citizens this can mean faster and more accessible services: from automated preparation of draft court decisions to more personalized electronic services in the "Diia" app. It is also a signal to investors and partners: Ukraine is building AI competencies — including a rise of 14 positions in the global AI development ranking, and recognition of Diia.AI in international competitions that reinforce this point.

First applications: courts and "Diia"

One of the priority directions is the judicial sphere. Models that will assist in preparing draft court decisions are planned to be tested — this could potentially shorten case processing times and reduce the routine burden on judges. However, automation in courts requires strict oversight, transparent algorithms, and appeal mechanisms.

For "Diia" they will create Diia AI LLM — a model focused on public services. Its integration is expected to speed up citizen services, but it will also raise requirements for security, personal data protection, and open testing of results.

How it will work

Beyond PL will provide computational resources and infrastructure expertise; models will be trained in secure environments and tuned to Ukrainian legal and administrative contexts. A key element is guarantees of retained control over data and algorithmic transparency.

Analysts note: technical cooperation must be accompanied by a legal framework, model audits, and open testing to minimize the risks of bias and information leaks.

Conclusion: what next?

This is a step toward technological autonomy, but not a magic solution. Investments, regulatory frameworks, independent audits, and workforce training are needed. Now the ball is in the partners' and Ukrainian authorities' court: will they turn this start into a systemic advantage for the security and quality of public services?