◆ AI Learning · Hard rule

How can a developer use text data without breaking copyright laws?

A developer must follow copyright laws when using text. Short phrases are usually not protected by copyright. Longer phrases need careful handling to avoid legal problems.

By Pawan Nayar · from the LLOS AI Learning record

The rule

A developer must follow copyright laws when using text.

Short phrases are usually not protected by copyright. Longer phrases need careful handling to avoid legal problems.

A developer should limit the length of text phrases. Phrases of nine words or less are often not protected by copyright. This is because they are seen as common language, not unique creative works. If a phrase is longer than nine words, a developer should shorten it. An ellipsis (...) can show that text was removed. This helps keep the data legally safe. Every piece of text should include where it came from. This means adding the movie name, lyricist, or actor. This practice supports fair use and gives credit to the original creator. Adding new meaning or context to text changes its purpose. This can turn entertainment data into educational or research data. This transformation can help with fair use arguments.

A developer creates a database of movie quotes. The developer ensures no quote is longer than nine words. If a quote is longer, the developer shortens it and adds '...'. The developer also includes the movie title and actor for each quote. This makes the database useful for study without copyright issues.

Words you'll need

Resolved from this lesson's own record — real definitions, not invented for this page.

Data Ethics

Doing the right thing with information about people

Natural Language Processing

AI understanding and generating human language

Information Extraction

Getting specific facts out of messy text

Synthetic Data Generation

Creating artificial datasets that imitate real data for AI

Questions people ask

More from the same record, pulled by shared topic — not written new for this page.

How should an AI model handle copyrighted slogans in its training data?

An AI model's training data must be cleaned of copyrighted slogans. This prevents the AI model from generating protected content.

How can an AI model avoid using copyrighted brand names and slogans from its training data in its output?

A developer must remove copyrighted brand slogans from the AI model's training data. The developer replaces brand names with a placeholder like `[brand]`.

How does an AI model protect confidential information and intellectual property?

An AI model uses a three-layer defense system. This system stops sensitive data from being shared.

How can an AI model avoid using copyrighted brand slogans?

A developer must check all slogans in the AI model's training data. This check happens in two steps.