LLMs and the Risk of Copyright Infringement

LLMs and the Risk of Copyright Infringement

The recent launch of Meta's Llama 2 has brought unprecedented interest in open-source large language models (LLMs). While the excitement surrounding this launch is understandable, it is crucial not to overlook the legal uncertainties surrounding intellectual property (IP) ownership and copyright in the generative AI space. Many assume that regulatory risks only concern the companies creating LLMs, but this assumption proves dangerous when we consider the poison pill of generative AI: derivative works.

What are derivative works?

A derivative work is a work that is based on or derived from another work. In the context of generative AI, a derivative work could be anything from a text that is generated by an LLM to an image that is created by an LLM.

Why are derivative works a problem?

The problem with derivative works is that they can be difficult to define. Under copyright law, a derivative work is any work that is "substantially derived from" another work. However, what constitutes "substantial" is often open to interpretation. This can create uncertainty for businesses that use LLMs to generate content.

For example, let's say a business uses an LLM to generate a text that is similar to a copyrighted work. Is this text a derivative work? It depends on how similar the text is to the copyrighted work. If the text is only slightly similar, it may not be considered a derivative work. However, if the text is substantially similar, it may be considered a derivative work.

The risks of derivative works

The uncertainty surrounding derivative works can create a number of risks for businesses. For example, businesses that use LLMs to generate content could be accused of copyright infringement if the content is considered to be a derivative work. This could lead to lawsuits and significant financial losses.

In addition, the uncertainty surrounding derivative works can make it difficult for businesses to license LLMs. Businesses may be reluctant to license LLMs if they are not sure what rights they are getting. This could stifle innovation in the generative AI space.

How to mitigate the risks of derivative works

There are a number of steps that businesses can take to mitigate the risks of derivative works. These steps include:

• Understanding the copyright law in your jurisdiction.

• Using LLMs in a responsible manner.

• Getting clear licensing terms from the LLM provider.

• Taking steps to protect your own IP.

By taking these steps, businesses can help to minimize the risks associated with derivative works and protect their IP.

The future of derivative works

The use of LLMs is likely to continue to grow in the future. As this happens, the legal challenges surrounding derivative works will become even more complex. It is important for businesses to stay informed about these challenges and take steps to mitigate the risks.

By understanding the risks of derivative works, businesses can help to ensure that they are using LLMs in a way that is both legal and ethical.