Open-ended debate

Upgrade to High-Speed Internet for only ₱1499/month!

Enjoy up to 100 Mbps fiber broadband, perfect for browsing, streaming, and gaming.

Visit Suniway.ph to learn

Copyright protects original literary and artistic works of natural persons.

But what about works generated through artificial intelligence?

Courts all over the world are still divided on this issue, especially when generative AI systems are trained on copyrighted works.

Generative AI is a type of artificial intelligence that creates new content, such as text, images, music, audio and video, by learning patterns from existing data. Powered by large-scale machine learning models, it generates original outputs based on user prompts rather than merely analyzing or classifying data. Like all AI, generative AI works by using machine learning models pre-trained on vast amounts of data.

An example of text generation would be creating articles, essays, reports and emails using tools like ChatGPT or Google Gemini. Or producing art, logos, or realistic images from text prompts using DALL-E or Midjourney.

Generative AI has several uses across industries, according to Amazon Web Services (AWS). For instance, financial institutions use chatbots to generate product recommendations and respond to customer inquiries. Lending institutions speed up loan approvals for financially underserved markets, especially in developing nations, while banks quickly detect fraud in claims, credit cards and loans.

In health care and life sciences, one of the most promising generative AI use cases is accelerating drug discovery and research with the pharma industry utilizing generative AI systems to generate and optimize protein sequences and significantly accelerate drug discovery while in the automotive and manufacturing sectors, they use generative AI technology from engineering to in-vehicle experiences and customer service.

In media and entertainment, from animations and scripts to full-length movies, generative AI models produce novel content at a fraction of the cost and time of traditional production. For instance, artists can complement and enhance their albums with AI-generated music while media organizations use it to improve audience experiences by offering personalized content and ads to grow revenues.

The same AWS article cited a report from Goldman Sachs, which showed that generative AI could drive a seven percent or almost $7-trillion increase in global gross domestic product and lift productivity growth by 1.5 percentage points over 10 years.

In another article, IBM explained that generative AI relies on deep learning algorithms that simulate the learning and decision-making processes of the human brain. These models work by identifying and encoding patterns and relationships in large amounts of data, then using that information to understand users’ natural language requests or questions and to respond with relevant new content.

It cited McKinsey research stating that one-third of organizations already use generative AI regularly in at least one business function. At the same time, industry analyst Gartner projects that more than 80 percent of organizations will have deployed generative AI applications or used generative AI application programming interfaces by this year.

The Organization for Economic Cooperation and Development revealed that the use of generative AI tools in 2025 is particularly high among students, with three-quarters of students aged 16 and over reporting use. Adoption is also widespread among those connected to the labor market, including employed individuals (41.1 percent) and the unemployed (36.7 percent).

But the use of generative AI tools to produce works across industries has attracted its fair share of criticism, especially from those who claim that their original works were used without their consent to train these learning models.

In a report, Reuters revealed that last year, US courts issued the first substantive, merits stage decisions addressing whether the use of copyrighted works to train generative AI systems constitutes fair use. And while these rulings do not settle all open questions and, in some respects, highlight emerging judicial disagreements, they represent a significant inflection point in copyright law’s response to large language models, image generators, and other foundation models.

In a Northern District of California court, it was held that Anthropic’s use of lawfully acquired books to train its large language models (LLMs) constituted fair use under the US Copyright Act. Here, Anthropic bought millions of print books, scanned and digitized them, and discarded the originals to build a central library for training its Claude AI models. The court emphasized the highly transformative nature of LLM training, saying that the models do not function as repositories of expressive content, but instead extract statistical relationships to enable new capabilities.

In another case, also in the Northern District of California, where 13 published authors used Meta for using their copyrighted books, sourced from online “shadow libraries” to train its Llama LLM, the court agreed that LLM training is transformative but emphasized that market harm remains the most important fair use factor and cannot be dismissed by analogy alone.

Fair use is a US copyright doctrine allowing limited use of protected material without permission, generally for purposes such as criticism, comment, news reporting, teaching, scholarship, or research, and is an affirmative defense against copyright infringement. Fair use is determined on a case-to-case basis through four factors: purpose and character of the use under which transformative of the work or adding new expression, meaning or message favors fair use; nature of the copyrighted work under which using factual work is more likely to be fair than using highly creative or unpublished work; amount and substantiality of the portion used so that using a small portion of the original is more likely to be fair and effect of the use upon the potential market of the copyrighted work.

Philippine laws adopt the same fair use doctrine under the Intellectual Property Code.

In one case, Thomson Reuters sued an AI-driven competition for using its Westlaw headnotes and key numbers in training Ross Intelligence’s AI-driven case search tool, and a US court ruled in favor of the plaintiff, finding that the headnotes were original and protected, and that Ross Intelligence’s use was not fair.

The debate has not been settled, but what is clear is that the use of generative AI is growing and is here to stay. But copyright laws must be able to evolve quickly to protect the rights of creators of literary and artistic works. One must not be sacrificed at the expense of the other.

For comments, email at [email protected]