Poland is rolling out a new artificial-intelligence feature inside mObywatel, its flagship government services app, positioning the move as a landmark step in digital administration across the European Union. Starting 31 December 2025, users will gain access to a virtual assistant powered by PLLuM, a Polish-developed language model trained on Polish-language public administration content. The Ministry of Digital Affairs says the goal is pragmatic: help citizens navigate official procedures faster, without turning the chatbot into a general-purpose “everything assistant.”

A chatbot built for bureaucracy—on purpose

According to Deputy Minister Dariusz Standerski, the assistant is meant to guide users through typical, high-friction administrative tasks—such as applying for a new ID card or passport, reserving a PESEL number, or reporting the sale or purchase of a vehicle. Instead of generating creative text or offering lifestyle tips, it is designed to answer questions quickly and point the user to the correct place: a form, a feature inside the app, or a relevant page on the gov.pl domain.

That “narrow” design is deliberate. The ministry emphasizes that the assistant does not interpret the law, does not replace professional advice, and does not function like a typical consumer chatbot that can draft letters, check the weather, or provide recipes. In other words, it’s an AI tool for administrative navigation—built to reduce confusion rather than expand into open-ended conversation.

Why PLLuM matters: language, context, and trust

The system’s backbone—PLLuM—has been presented as a model built around Polish language and Polish institutional realities, trained on government texts so it can “understand” the way Polish administration communicates. In the ministry’s framing, that also supports a broader political and technological ambition: greater control over critical digital infrastructure and reduced reliance on foreign AI systems when dealing with public services.

Standerski described the rollout as the first time a government app in the EU offers such an assistant powered by a domestic model. It’s a bold claim, and it clearly signals that Warsaw wants the project to be read not only as a convenience feature but also as a statement about technological sovereignty.

“No personal data” — and strict access limits

A core promise attached to the launch is privacy. The ministry states the chatbot does not process users’ personal data and has no access to such data. Just as importantly, it also has no access to state registers, meaning it cannot “look you up,” fetch your records, or make decisions based on sensitive databases. This is meant to keep the assistant in an informational and navigational role rather than making it a gatekeeper of official outcomes.

Who gets it, and where it appears in the app

The assistant will be available to users who have mObywatel version 4.71.1 or newer. It will not be available in mObywatel Junior. In the interface, users are expected to find it near the notifications area (top right) and in the “More” menu, opening a chat window where they can type questions and receive guided responses.

From pilot to full rollout

Before the full launch, the assistant was tested in a pilot lasting nearly two months, involving roughly 5% of users, according to the deputy minister. Feedback from this period reportedly helped refine the system before nationwide deployment.

A political aftershock: AI, TikTok, and disinformation

The announcement also came with a warning about AI used for manipulation rather than public service. Standerski referenced a TikTok account accused of pushing pro-“Polexit” content using AI-generated personas, and said the ministry had been coordinating reporting efforts with the platform. He also stated he had contacted Henna Virkkunen, a European Commission vice president responsible for technological sovereignty, urging the Commission to examine whether TikTok fulfilled its obligations around risk assessment and response to disinformation.

In effect, the ministry is presenting two sides of the same coin: on one side, AI as a tool that can make the state easier to use; on the other, AI as an amplifier of deception—especially when synthetic content is not clearly labelled and targets younger audiences.

What this rollout signals for digital government in the region

Whether the assistant becomes a genuine quality-of-life improvement will depend on mundane but crucial factors: accuracy, clarity of routing users to the right procedures, and the ability to handle real-world questions without hallucinating. Yet even at launch, the political significance is hard to miss. Poland is attempting to show that a member state can deploy AI in public services while drawing bright lines around privacy, register access, and scope—framing the project as both practical modernization and a step toward domestically anchored digital infrastructure.