A team of researchers investigated which language artificial intelligence understands best, with Polish emerging as the clear leader. Out of 26 languages tested, English ranked only sixth, according to a joint study by the University of Maryland (UMD) and Microsoft.
“Our findings were both surprising and counterintuitive,” the researchers wrote. “English did not dominate across all models. It finished sixth in long-text evaluations, while Polish led overall.”
The researchers evaluated how major AI systems—including OpenAI, Google Gemini, Qwen, Llama, and DeepSeek—handled identical prompts in 26 languages. Polish achieved an average accuracy rate of 88% in task completion, outperforming every other tested language.
Polish Emerges as AI’s Most Command-Friendly Language
The analysis showed that AI systems followed instructions most accurately in Polish. “It appears that Polish delivers the clearest commands to artificial intelligence,” the Polish Patent Office noted in a Facebook post. “People find it difficult to master, but AI does not.”
AI models displayed an unexpected strength in understanding Polish, despite limited Polish-language data compared to English or Chinese. The findings suggest that AI reasoning may depend less on data volume than on language structure.
In contrast, Chinese performed poorly, ranking fourth from the bottom among all tested languages. Researchers found that even smaller linguistic datasets could yield higher AI comprehension if the language’s structure aligns well with machine learning models.
Top Languages for Effective AI Communication
The study identified the ten most efficient languages for interacting with AI systems:
- Polish – 88%
- French – 87%
- Italian – 86%
- Spanish – 85%
- Russian – 84%
- English – 83.9%
- Ukrainian – 83.5%
- Portuguese – 82%
- German – 81%
- Dutch – 80%
The results revealed that AI responds most precisely to structured, inflection-rich languages like Polish and French, challenging long-held assumptions about English’s dominance in the digital world.
