This website uses cookies so that we can provide you with the best user experience possible. Cookie information is stored in your browser and performs functions such as recognising you when you return to our website and helping our team to understand which sections of the website you find most interesting and useful.
Open-source and AI
May 09, 2023
In recent months, a great deal has been written and discussed about Artificial Intelligence (AI), its advantages, and the questions it raises. From the aspect of copyright law, many questions arise regarding the lawfulness of using copyright-protected works to train AI programs, the works such programs produce after their training is complete, and whether using such programs has the potential to infringe the copyright of others.
On top of those questions, there are also contractual questions, such as which software license those AI programs operate under and the impact of that license’s terms on the users, especially when some of these programs are distributed under one of the Open-Source licenses that place certain limitations on how a produced work can be used.
An Open-Source software package is a software whose source code is freely available. The code is easily accessible on the Internet for anyone to download, review, adapt, and/or deploy. However, commercial use of Open-Source code can entail legal risks. Some Open-Source software licenses require companies to use the code to make their own proprietary code public in some circumstances.
To the extent that an AI program is distributed under an Open-Source license (such as the GPL license), which obligates users modifying the source code to publicly disclose their source code, further questions arise. For example: what is the status of a software product that includes trained AI models? Should the training be regarded as a “modification” of the open-source code that must be disclosed to the public?
In addition, another question arises regarding how such software can interface with other software in a way that does not require exposing source code that has been independently developed. Will linking, using a dynamic or static library, necessitate the making public of the developer’s code? And what architectural decisions ensure that interactions between two software do not create legal risks?
These questions may be answered by our firm’s team, which includes attorneys specializing in copyright, as well as patent attorneys and software experts.
Our team can assess the commercial risks resulting from different Open-Source licenses while taking into consideration the particular use cases of Open Source software in the organization’s products and services and provide ongoing responses to development teams and business managers in special cases of Open Source code use.
This article is provided for general information only. It is not intended as legal advice or opinion and cannot be relied upon as such. Advice on specific matters may be provided by our group’s attorneys.