Session 6: AI & Communication
Natural Language Processing is how AI learns to understand and use human language - the words we speak and write.
AI can't understand "hello world" as words like you do. First, it must break language into pieces called tokens.
Sentence: "Can't believe it's Friday!"
After tokenizing, AI converts each word into a series of numbers. This is called a word embedding.
The biggest challenge: the same word means different things in different contexts!
Modern AI models look at surrounding words to understand which meaning is correct!
This is where AI fails most often! Humor requires understanding what ISN'T said.
Not all languages work the same way! Direct translation often fails.
Language AI learns bias from the text it's trained on - which reflects human prejudices.
AI understands language through numbers and patterns, but still struggles with human complexity
Next Session: AI & Privacy