The Mozilla Foundation recently unveiled its "Rise25" list, spotlighting 25 individuals leading the next wave of AI innovation. I am honored to be among them.
The selected honorees are united by a shared commitment to equity, applying the open source ethos across various fields to tackle issues like racial, gender, and economic inequality. Whether in technology or business, the AI solutions proposed by these individuals are deeply rooted in this principle.
Open source platforms are akin to public libraries—places where everyone has equal access. Leveraging the open-source model is essential to counteract the growing threat of generative AI being monopolized by a few tech giants and to encourage the sharing of resources.
Why is this crucial? As AI models become larger and require more data—much like how a larger brain needs more books to read—the cost of training these models rises steeply, often becoming feasible only for a few large tech companies. By open sourcing foundational models, various industries can pool their data and tailor AI systems to meet specific needs, resulting in solutions better suited to real-world challenges.
For example, one of this year's Rise25 honorees, Peter Gault, demonstrates the power of collaboration with educators through his open source project, Quill.org. Gault's platform leverages AI to provide free tools for students to improve their reading and writing skills while offering teachers high-quality curriculum content. This helps level the playing field, giving students with limited resources greater opportunities to learn.
Quill collaborates closely with K-12 teachers to develop the content for classrooms. The platform's Teacher Advisory Council (TAC), composed of over 300 members, tests every learning activity before it is released, using real-world scenarios to guide AI development. This ensures that AI-generated feedback is relevant to the needs of teachers and students, making it a practical and effective tool. In contrast, many generative AI systems created by tech teams tend to offer direct answers, which doesn't always address the underlying learning gaps of students.
Currently, Quill's services are available only in English. In Taiwan, there are also ongoing efforts to develop open source models, datasets, and evaluation systems. One standout example is the "TAiwan Mixture of Experts (Project TAME)," a collaborative initiative involving experts from various fields such as chemical engineering, electronics, healthcare, content creation, and law, alongside a team from National Taiwan University.
In the educational sphere, open source initiatives like the Awesome Taiwan Knowledge (ATK) cover a broad array of subjects, including history, culture, politics, education, and current affairs. These projects are curated and validated by educators and subject matter experts, ensuring that AI tools align with local contexts and remain accessible to the general public.
While debates continue about whether generative AI is just a passing trend, those dedicated to the open-source approach are steadily solving previously insurmountable challenges, paving the way for a more equitable and inclusive future.
(Interview and Compilation by Hsin-Ting Fang. License: CC BY 4.0)