AI Chatbots Fuel Surge in Complaints and Filings to Polish Offices

Polish government offices and customer service departments are seeing a growing number of letters, applications and complaints written with the help of artificial intelligence. The phenomenon is spreading fast enough that officials and lawyers are already calling it a snowball effect: a single prompt typed into ChatGPT or Gemini can generate a multi-page document in seconds, packed with references to regulations that still have to be carefully verified.

Where the Surge Comes From

The mechanism is simple. A consumer who wants to file a complaint about a product or service no longer needs to know the law or hunt for a template online. It's enough to describe the situation to a chatbot, which within moments produces a ready-made justification citing specific articles of the Consumer Rights Act or the Civil Code.

The same mechanism is at work in dealings with the administration. Residents are sending more freedom-of-information requests to the city offices in Wrocław and Kraków, and regional chambers of audit are receiving more complaints and inquiries about local government actions. These documents share a distinctive form: they are longer than before, follow a repetitive structure, and are heavy on bold text and extensive citations of case law.

The Cost for Businesses

For businesses, the problem isn't the sheer number of letters but the asymmetry of the work involved. Writing a complaint takes a consumer literally a moment, while verifying it properly on the company's side requires the time of a customer service employee, and sometimes a legal department too, even if most of the arguments turn out to be groundless.

AI has significantly lowered the cost of building an argument on the client's side, but not to the same degree on the verification side for businesses - Kinga Konopelko, attorney-at-law

On top of that there's a psychological effect lawyers describe as "what's the harm in trying." The low barrier to entry means clients now file formal complaints in cases they would have previously shrugged off, because drafting the letter itself used to seem too time-consuming or too legally complicated.

Hallucinations in Legal Filings

The problem isn't just volume, it's the quality of the content. Language models can cite court rulings or official positions that don't actually exist, or apply legal grounds that don't fit the situation described. A chatbot can also reinforce a client's mistaken belief that their rights were violated, even when it was given incomplete or imprecise information.

The Energy Regulatory Office found something in some of the submitted documents that gives away their origin even more clearly than the writing style: leftover fragments of prompts, meaning the instructions the sender typed into an AI tool, which were mistakenly left in the final version of the document sent to the office.

How Institutions Are Responding

The administration is starting to fight back with its own artificial intelligence. The Regional Chamber of Audit in Opole purchased a tool meant to help assess whether a given filing was prepared by AI, aimed at speeding up initial screening and letting officials focus on the substance of a document rather than its length.

At the same time, offices stress that how a document was prepared doesn't exempt them from the obligation to process it. Every application, whether written by a person or a chatbot, must be handled according to the applicable regulations, which means the growing volume of letters is placing a real burden on officials' working time.

For Polish companies and public institutions, this means rethinking how they handle complaints and applications before the scale of the phenomenon grows even further. Experts point out that some AI-generated complaints do flag real errors in terms of service, offers or procedures that had previously gone unnoticed, so automatically dismissing such letters as spam would be a mistake.

The trend also has a time dimension that explains the term "snowball effect": the more people see that an AI-written complaint or application gets results, the more eager other customers and residents are to use the same tool, which in turn fuels further growth in the number of letters reaching companies and offices.