Andriy from Kharkiv submitted a name in an online form — one of more than 3,000 suggested options. His idea did not make the final, but he voted. So did 135,999 other Ukrainians. The winning option was 'Syaivo' («Сяйво») — 22,601 votes out of more than 136,000 cast.
This is not just a rebranding of another chatbot. 'Syaivo' is the first national large language model (LLM) being jointly developed by the Ministry of Digital Transformation and Kyivstar. The Ministry positions the project as a foundational digital infrastructure for the country — on a par with the power grid or roads, only for data.
The scale of the ambitions is commensurate: the model is planned to be integrated into government services, educational platforms, business solutions and — it is worth emphasizing separately — defense systems. In other words, this is an AI that will process sensitive data of millions of people and, potentially, influence decisions in the security sphere.
An open vote for the name is an atypical step for Ukraine toward public involvement in a state tech project. It creates a sense of co-authorship. The problem is different: a public technical standard for the model, an audit framework, or a mechanism for independent verification — have not yet been published. The name was chosen together, the architecture was not.
Kyivstar as the private partner of the state in a project of this scale is also a question without an answer. Who controls the data that 'Syaivo' will process? What security protocols are envisaged for defense applications? For now, the Ministry of Digital Transformation has limited itself to announcing the voting results.
A national-scale LLM is not a product you can launch and see what happens. It is an infrastructural solution whose mistakes are costly and slow to fix. Countries that built similar systems without transparent standards from the start — India with Aadhaar, Estonia with X‑Road in its early versions — spent years patching vulnerabilities after scaling.
'Syaivo' has received a name. The next test is whether it will receive a public technical regulation before it is integrated into systems where the cost of failure is measured not in App Store ratings.
