A company develops an AI system for automatic translations.

It is trained with millions of texts in different languages.

In everyday use the translator works surprisingly well.

Many people use it for emails, websites, and conversations.

One day the AI translates a simple sentence in a very strange way.

The translation sounds grammatically correct but makes no sense.

The original sentence was actually easy to understand.

Yet the AI suddenly produces a completely wrong meaning.

Question:
Why can an AI produce a grammatically perfect sentence that is still completely wrong?

Solution will be unlocked tomorrow.