Abstract
While broad and multi-modal Artificial Intelligence (AI) has been a subject of discussion for decades, it has only recently become a reality. This raises concerns about power distribution, accountability, and ethical considerations within the AI landscape. In 2020, 31 scientists conducted a study that had appeared in a prominent journal. The study described successful trials of an AI that looked for signs of breast cancer in medical images. However, according to its critics, the research team provided so little information about its code and how it was tested that the study promoted proprietary tech that was protected by trade secrets. The lack of transparency in the legal aspects prevented the community from examining and replicating the publication. In today’s world, AI researchers working for the industry labs of a handful of global corporations have access to real-world problems and large amounts of data that is not publicly available. They have access to vast amounts of resources such as compute power but they are also known to employ vast and cheap resources in lower and middle income countries for doing data labeling or the current alignment methods, like Reinforcement Learning from Human Feedback. And as always with these industry labs, they are driven by economic incentives.
Moreover, now 3 years later, after having witnessed the birth of new open-sourced AI communities, a leaked document from an engineer mentions that a company cannot directly compete with open source and that open source companies are doing things in weeks with minimal cost and parameters that are challenging for proprietary companies.
As we delve into this debate, we will focus less on the current possibilities of medical AI, but explore the implications of various ownership and control models for AI. First, we will analyze historical and literary precedents, examining the tensions between open-science and open-source collaboration and proprietary control. However, we also look into the impact of historical decisions such as the event that took place on 26th June 2000, where US President Bill Clinton and UK Prime Minister Tony Blair presided over a carefully choreographed piece of scientific theater and announced that the human genome should be used freely for the common good of the whole human race, thus open sourcing it.
By delving into these considerations, we aim to foster an insightful panel discussion that helps shape the future of AI ownership and control, aiming to maximize benefits while minimizing potential risks and inequalities. It echoes the struggle for intellectual property rights, the balance between collaboration and inclusive progress, proprietary interests, and the impact of technological advancements on global societal structures.