Trademarks, volunteering, and code-generating LLM
This year's LibrePlanet saw a large swath of fantastic talks on a wide variety of topics related to free software, including a track specifically dedicated to licensing. This track featured three talks: one about the free licensing of trademarks, another hosted by the FSF's Licensing and Compliance Lab, and a post-LibrePlanet presentation on the ethics of code-generating large language model (LLM) systems.
On the first day of LibrePlanet talks, Julian Daich from LINDS, a 501(c)(3) nonprofit that is developing biomedical technology released under free licenses, presented an interesting approach to applying free licensing to trademarks. As you may already know, trademarks are used to identify services or products as authentic to a specific organization. Daich is working toward a system of trademark licensing that enables the identification of free works by a free trademark. This system includes at the same time the licensing for the commercial use of the trademarked services or products. The hope is to offer a method for uncompromisingly protecting the freedoms of work created by a concerned community, while offering a system for that same community to generate revenue via commercial (but still free as in freedom) use and distribution under different branding.
Closer to the end of the day, the licensing volunteers of the FSF's Lab gave a talk about their history with free software, as well as the experiences and benefits of working as licensing volunteers.
Yoni Rabkin, who facilitated the talk and has been a volunteer since 2006, expressed how he identifies with the ideals of the free software movement, and how the combination of his passion for free software and legal work made being a licensing volunteer a perfect fit.
Panos Alevropoulos, former FSF intern and most recent addition to the team of licensing volunteers, who likes to go by "Panos," described the importance of informing the public about software licensing, invoking the concept of "legal awareness." Legal awareness is when everyday people are empowered with basic literacy in common law. When legal awareness is achieved, the public can apply this knowledge to better participate in a democratic society, becoming informed and active citizens. In today's technology-pervasive world, it is important to create awareness about free software and the laws surrounding it. Panos also commented on the Free Software Directory as a robust resource for assistance in free software licensing and cataloging of free programs.
Paulius Gaulubickas, another licensing volunteer, touched on the need for an international body to facilitate arbitration for enforcing free software licenses on a global scale, especially for those without access to legal assistance.
Yuchen Pei and Ineiev were unable to attend LibrePlanet 2023, but I'm honored to mention them here for their great contributions as volunteers to the Lab.
And as for me, I was able to describe the time when I first discovered the free software movement, which has provided me a vehicle to express myself in order to make the world a better place through free software. To watch a video of the complete talk, please go to: https://media.libreplanet.org/u/libreplanet/m/volunteering-for-the-licensing-and-compliance-lab/.
A few weeks after the in-person conference, Harm de Vries, a staff research scientist at ServiceNow, and Leandro von Werra, a machine learning engineer at HuggingFace, represented the BigCode Project and gave an online presentation on one of the hottest topics of today, the ethical implications of code-generating LLMs. The BigCode project is a scientific collaboration (with over 350 participants) working on the responsible and ethical development of code-generating LLM systems. In their talk, they discussed how they navigate the legal, ethical, and governance-related aspects around the development of these models, including how to develop a permissively licensed code dataset. They also feel it is essential to give developers the option to remove their code from the training data, redact personally identifiable information (PII), and attribute generated programs to the original code snippet. This workshop has not been published yet, but stay tuned to the LibrePlanet 2023 talks page for its eventual release.
On Sunday, the Lab hosted a licensing table at LibrePlanet. The intention was to provide a space for an open discussion about free software licensing, where conference attendees could drop in, either virtually or in person, and ask questions. I enjoyed the conversation with attendees and was delighted to see their interest in free software licensing. We hope to expand the licensing table in the years to come.
LibrePlanet continues to be a magnet for innovations in ethical licensing of which I am proud of being a part, and I already look forward to LibrePlanet 2024. I hope that you will watch the recorded talks, and I look forward to seeing you at a future conference, whether online or in person.