Researchers from the University of Maryland and Microsoft found that Polish ranks highest in AI comprehension, while English placed sixth. The study examined how 26 languages perform when used to prompt artificial intelligence systems.
“Our results revealed surprising outcomes,” the authors explained. “English did not perform best across all models. It ranked sixth, while Polish led with the strongest results.”
The team tested leading AI models, including OpenAI, Google Gemini, Qwen, Llama, and DeepSeek, using identical prompts translated into 26 languages. Polish achieved an average accuracy rate of 88% when completing assigned tasks.
AI Excels at a Language Humans Find Complex
Researchers concluded that Polish delivers the most precise instructions to AI systems. The Polish Patent Office noted, “AI understands Polish far better than most humans. Although Polish is known as one of the hardest languages to learn, artificial intelligence handles it effortlessly.”
Despite limited Polish-language training data compared with English or Chinese, AI models still showed exceptional proficiency in Polish. By contrast, Chinese performed poorly, ranking fourth from the bottom among all languages tested.
Top Languages for AI Performance Revealed
The study ranked the ten most effective languages for conversational AI as follows:
- Polish – 88%
- French – 87%
- Italian – 86%
- Spanish – 85%
- Russian – 84%
- English – 83.9%
- Ukrainian – 83.5%
- Portuguese – 82%
- German – 81%
- Dutch – 80%
Researchers believe these findings could reshape how developers design and train AI models, challenging long-held assumptions about English dominance in artificial intelligence communication.
