AI Brain Alignment Depends on Training Language, Not Language Itself
Researchers found that AI models align more closely with human brains when trained on the same language. This suggests language-specific training data shapes how AI understands us, not an inherent property of any language.
Researchers from the University of California released a study showing that AI models align more closely with human brains when trained on the same language. The team used fMRI data from 112 participants speaking English, Chinese, and French, comparing it to seven different AI models. They discovered that the alignment depends on the language used to train the AI, not the language itself.
This matters because it shows that AI models trained on diverse languages will understand and communicate with people more effectively. For example, an AI trained on Chinese will align better with Chinese speakers' brains than an English-trained AI. This could lead to more personalized and accurate AI assistants for different language groups.
If you're curious about how AI understands language, try asking a multilingual AI assistant like Google Translate or DeepL a question in your native language. Observe how it responds and compare it to how you might explain the same concept to a friend. This simple exercise can help you see the differences in language understanding firsthand.