TL;DR — Key Takeaways
- The Eclipse Foundation is forming the Sovereign AI Foundation as a vendor-neutral initiative focused on helping organizations understand and manage dependencies across open source AI platforms.
- The initiative does not plan to build software or specifications. Instead, it aims to foster an ecosystem of independently governed technologies that can address areas such as security, ethics and transparency.
- Sovereign AI can encompass everything from retaining control of data to running workloads on locally operated infrastructure, making sovereignty a spectrum rather than a single technical requirement.
The Eclipse Foundation today announced it is forming a Sovereign AI Foundation, a vendor-neutral initiative that will help the open source community navigate issues as organizations look to build and deploy artificial intelligence (AI) platforms that they control and maintain.
Founding members of the Sovereign AI Foundation include CEA LIST, EclipseSource, Engineering Ingegneria Informatica, Ericsson, Eurotech, Infosys, Kentyou, KU Leuven, Open Elements, Red Hat, Renesas Electronics, Robert Bosch, Thales, The IO Foundation, TypeFox, the University of York, and Vector Informatik.
Specifically, the Sovereign AI Foundation will focus on building solutions based on the member input and enabling organizations to navigate the dependencies that exist between open source elements of an AI platform, says Michael Berns, head of AI for the Eclipse Foundation.
At the same time, however, the Sovereign AI Foundation has no plans to develop software or specifications. The goal is to create an ecosystem of independently governed open source technologies that address different requirements, such as ethics and security, but work together when appropriate, notes Berns. “We want to increase transparency while also reducing risks created by dependencies,” he says.
There is already no shortage of AI consortiums, but this latest one from the Eclipse Foundation appears to be mainly focused on fostering a dialogue rather than helping members market specific offerings that are made available under some type of open source license. The degree to which those interactions at this early stage will benefit organizations deploying AI workloads is unclear but, in general, AI sovereignty is an outgrowth of the rise of digital sovereignty initiatives through which organizations are attempting to reduce their dependency on IT platforms that are subject to regulations defined by a foreign government. The European Union (EU) in particular is keenly interested in promoting digital sovereignty across its member bloc of 27 nations.
Of course, there are multiple levels of sovereignty. In some cases, organizations only want to maintain independent control of their data. In other instances, all the IT infrastructure running a workload needs to be deployed in a data center within a specific geography in a way that is managed locally by IT personnel who are citizens of a specific country.
Each organization will, naturally, need to determine what level of sovereignty is achievable and maintainable, especially if they rely on servers and other IT infrastructure that has components that were manufactured in China. In the meantime, however, some level of independence is always advisable given rising geopolitical tensions, notes Berns.
Ultimately, it’s impossible to make an IT decision that at some level isn’t impacted by regulatory requirements. The challenge and the opportunity now is to, arguably, reduce the number of instances where regulations imposed by multiple countries might apply. Unfortunately, at this stage that may be easier said than actually done. Nevertheless, independence is rarely won in a single day, so in most instances, AI sovereignty may require a significant amount of patience and forbearance over multiple years before being actually achieved.
