Science & Technology · 5 Intermediate · September 2, 2026

AI Programs Need Training in Languages

Photo: Christina Morillo via Pexels

Listen

Vocabulary

artificial intelligence /ˌɑr.tɪ.ˈfɪʃ.əl ɪn.ˈtel.ə.dʒəns/ noun

computer systems that can think and learn like humans

Artificial intelligence can answer questions and write emails.

training /ˈtreɪ.nɪŋ/ noun

the process of teaching a system by giving it examples and data

The AI received training in English by reading thousands of books.

limitation /ˌlɪm.ɪ.ˈteɪ.ʃən/ noun

something that restricts or prevents; a weakness

A major limitation of early AI was that it only worked with one language.

pattern /ˈpæt.ərn/ noun

a regular arrangement or repeated order of things

The AI learned the pattern of how English words usually appear together.

data /ˈdeɪ.tə/ noun

information or facts, usually in numbers or organized form

We need more data in African languages to improve AI systems.

struggle /ˈstrʌg.əl/ verb

to have difficulty doing something

The AI struggled to understand the question because it was in a language it never learned.

Article

Artificial intelligence tools like ChatGPT can answer questions and write text. However, these AI systems have an important limitation. They can only speak and understand languages that humans taught them during training. If an AI system learned English and Spanish, it cannot suddenly speak French or Japanese without that training.

How AI Learns Languages

AI systems learn languages by reading millions of texts written by people. Engineers feed the AI large amounts of documents, books, and websites in different languages. The AI studies patterns in these texts and learns how words connect together. This process is called training. After training, the AI can recognize and produce language it has seen many times before.

The Limitation Problem

If someone asks an AI system a question in a language it has not learned, the AI cannot answer well. For example, an AI trained only on English texts will struggle with a question in Korean. The AI simply does not have the knowledge or patterns to understand Korean words.

Why This Matters

This limitation affects people around the world. Many languages have less training data available online. Languages spoken by fewer people may not have enough texts for AI systems to learn. This means speakers of smaller languages cannot easily use AI tools.

Building Better AI

Companies are working to improve this situation. They collect more training texts in different languages. They also create AI systems designed specifically for certain languages. In the future, AI may become available in more of the world's languages.

Discussion Questions

What languages do you speak? Do you think AI should learn your language?
Why do you think some languages have more training data than others?
How could the world benefit if AI worked in all languages?
Have you used an AI tool like ChatGPT? What language were you using?
Do you think it is fair that AI works better in some languages than others?

Open the full interactive lesson with tap-to-hear on every word →

Ready to speak English with real confidence?

Work 1-on-1 with Digby R. Kerr.

Claim Your Free 30-Min Class