Register here!
Note: the event is fully online, and attendance is free. Please register, and you will receive Zoom link one day before the event.
Dr. Sarah Myers West, AI Now Institute
In light of nearly all major AI labs releasing their own open-weight models over the last couple of years, it is perhaps the perfect time to talk about AI and Open Source. This talk will focus on the notion of "open" AI: what exactly is "open" AI, and does it truly exist? We will examine the discourse around "open source" and "open science" in the context of AI industry, and how (mis)understanding of these concepts, when applied to AI, shapes businesses', the public's, and policymakers' knowledge of AI capabilities and limitations.
The development of "open" AI benefits from the efforts of a diverse community of creators, both for-profit and non-profit: from major AI labs, to research institutes and universities, and finally OS contributors. We will discuss the resources and components an AI creator needs to build AI from scratch and demonstrate the disparity in access to such resources across different types of AI creators. Taken together, the scale of resources required to build and deploy AI and intensifying competition beg the question: why have all major for-profit AI creators (with the exception of Anthropic, for now) embraced an "open" strategy? We will provide an overview of the main goals of this strategy, likely future moves toward these goals and actions already taken to achieve them. This will enable us to better understand the impact of "going open" on the AI industry ecosystem and assess risks, as well as provide relevant context to analyze existing regulatory efforts.
The development of "open" AI benefits from the efforts of a diverse community of creators, both for-profit and non-profit: from major AI labs, to research institutes and universities, and finally OS contributors. We will discuss the resources and components an AI creator needs to build AI from scratch and demonstrate the disparity in access to such resources across different types of AI creators. Taken together, the scale of resources required to build and deploy AI and intensifying competition beg the question: why have all major for-profit AI creators (with the exception of Anthropic, for now) embraced an "open" strategy? We will provide an overview of the main goals of this strategy, likely future moves toward these goals and actions already taken to achieve them. This will enable us to better understand the impact of "going open" on the AI industry ecosystem and assess risks, as well as provide relevant context to analyze existing regulatory efforts.