Skip to content
BENIBIX Learn

NLU vs NLG

People think NLU and NLG are two names for one thing, since both deal with plain language. They point in opposite ways. NLU reads words and turns them into meaning a program can use, like a goal or a date. NLG starts from data or a goal and writes words for a person to read. A voice assistant does NLU to hear you and NLG to answer.

AI🔤 NLP & Language

Natural Language Understanding (NLU)

The part of language AI that finds what text means. Then a program can act on it.

Natural language understanding, or NLU, is the part of NLP that works out what a piece of text means. It turns words into something a program can act on. That might be a goal, a label or a list of facts. A request to book a table for two at 8 becomes a booking, with a time and a count of people. Finding the mood of a text or the names in it are NLU tasks too. It is the reading half of language AI.

Example
Lucy types 'cancel my 3 pm meeting' into her calendar app. NLU finds the goal, cancel, and the thing to cancel, the 3 pm meeting.
Why it matters
A program can only act on words once it knows what they mean. NLU is the step that gets it there.
Full entry
AI🔤 NLP & Language

Natural Language Generation (NLG)

The part of language AI that writes. It turns data or ideas into text.

Natural language generation, or NLG, is the part of NLP that writes text. It starts from data or a goal and ends with sentences a person can read. Older systems filled in templates, like 'Sales rose by 5%'. Today, large language models write most of this text from scratch. Weather reports, game recaps and chatbot replies are all NLG. It is the writing half of language AI.

Example
A weather site turns its numbers into a line of text. It says rain comes after 4 pm, with a high of 12 degrees. No person wrote it.
Why it matters
It lets a program report to people in words, not just tables. That is how most AI output reaches us.
Full entry

In one line

Natural Language Understanding (NLU)
The part of language AI that finds what text means. Then a program can act on it.
Natural Language Generation (NLG)
The part of language AI that writes. It turns data or ideas into text.