An undisclosed AI-written government report cost an official her job in 24 hours

What happened

Dr. Barbara Mróz-Gorgoń, Poland's government plenipotentiary for promoting the Polish brand, published a report titled "Polaków portret własny" (Poles' Self-Portrait) through the Marka Polska Think Tank foundation. On August 20, 2026, critics publicly flagged factual errors and internal contradictions suggesting extensive AI generation - one questioned whether the Ministry of Sport and Tourism had paid 200,000 PLN for an AI-written document.

Mróz-Gorgoń acknowledged using AI, describing the report as an internal working draft that should never have become public. The Ministry disputed that the report was part of any contract, clarifying its actual agreement covered different work. Facing calls for dismissal and a threatened complaint to anti-corruption authorities, she resigned the next day, August 21, 2026.

The business problem

A document produced with undisclosed AI assistance was treated internally as a finished, credible deliverable, and released without anyone checking whether it could withstand public scrutiny.

What could have helped

• Undisclosed AI use in a public-facing or official document is a transparency risk independent of whether the content happens to be accurate
• A document's internal label - "just a working draft" - doesn't protect it once it's released; treat anything you create as if it will become a finished, public document
• Organizations need a clear internal rule for when AI-assisted work requires disclosure or additional review before release
• Reputational consequences from an AI-transparency failure can move faster than any formal investigation - this one played out in about 24 hours

Where this points