Open source technology has a long history of being a counterbalancing force to closed, proprietary systems. For decades it was open source versus corporations. Then Microsoft (closed, proprietary) bought GitHub (most used repository of open source code). Today, in the AI battles, Facebook/Meta, IBM and Oracle, along with universities and the National Science Foundation, announced the AI alliance - dedicated to open AI models. This is part of the larger debate about building responsible/trustworthy/safe/ethical AI.
So some of the world's biggest tech companies, many who have thrived on proprietary, patented, trademarked and close source code, are now arguing that an open community of developers is the way forward to protect us from the harms of AI.
This is one more step in both the commercial battles for market dominance and the definitions of words such as safety, ethical, trustworthy and responsible (in the context of AI.) For example, effective altruists and longterm (ers) (ists) focus on the word "safety." They're bogeyman is the potential for AI to destroy humanity. This group, the AI Alliance, uses the terms "open" and "responsible." They're bogeyman appears to be the other companies who've already launched proprietary models - like Google and Microsoft.
The mix of organizations and funding in these AI debates includes corporations, governments, and numerous nonprofits - not only universities, but also groups of developers and advocacy organizations. Philanthropic funding is very much in the mix. The direction of AI development is not simply an external force acting upon the nonprofit/philanthropic sector; it is being shaped by numerous actors within the sector. The meaning and purpose of "open" in this context is neither static, nor simple.
No comments:
Post a Comment