Researchers discovered that artificial intelligence understands Polish better than any other language, while English ranked only sixth. A study by the University of Maryland and Microsoft compared 26 languages to identify which ones AI models process most effectively.
“Our experiment produced unexpected results,” the report’s authors explained. “English did not dominate across models. When analyzing longer texts, it placed sixth, while Polish performed best.”
The team tested leading AI systems from OpenAI, Google Gemini, Qwen, Llama, and DeepSeek. Each model received identical prompts in 26 languages, allowing researchers to measure accuracy and consistency. Polish achieved an average accuracy of 88% in completing tasks, outperforming all others.
AI Excels in a Language Humans Find Difficult
According to the Polish Patent Office, “Polish proved the most precise for giving commands to AI. Ironically, humans struggle with it, but AI does not.” Despite limited Polish-language data available for training, AI models still demonstrated remarkable fluency and comprehension.
This outcome surprised researchers, as English and Chinese typically dominate AI training datasets. Yet Chinese performed poorly, ranking fourth from the bottom among all tested languages. The findings challenge long-held assumptions about which languages AI systems handle best.
Ranking the Top Languages for AI Communication
Researchers ranked the ten most effective languages for conversational AI based on overall accuracy:
- Polish: 88%
- French: 87%
- Italian: 86%
- Spanish: 85%
- Russian: 84%
- English: 83.9%
- Ukrainian: 83.5%
- Portuguese: 82%
- German: 81%
- Dutch: 80%
The study shows that AI’s linguistic strengths differ from human expectations. While humans view Polish as complex, AI models interpret it with exceptional clarity.
