Breaking Barriers: The Game-Changing Global Alliance Revolutionizing AI Development through Open Source Technology
Published on: March 10, 2024
On December 5, a landmark initiative in the AI industry was announced with the launch of a global alliance for AI open source development. This alliance, led by Meta and IBM, includes over 50 organizations such as Intel, Oracle, Dell, AMD, Sony, Hugging Face, and Stability AI. Its formation is seen as a direct challenge to the dominance of closed AI models developed by companies like OpenAI, Google, and Anthropic, which have been central to AI discussions throughout 2023.
The alliance boasts significant participants from outside Silicon Valley, including NASA, Cleveland Clinic, and the European Organization for Nuclear Research (CERN). It also includes academic institutions such as UC Berkeley, Yale, and universities from India, Japan, the U.K., and the UAE's Mohamed bin Zayed University of AI. This diverse membership underscores the alliance's commitment to a global perspective on AI development.
Positioning itself as a consortium of safety-conscious AI developers, the alliance aims to 'better reflect the needs and complexity of our societies.' It seeks to quickly identify and mitigate specific AI risks, positioning itself between the extremes of those dismissing AI guardrails and proponents of strict regulation. The alliance has announced its intent to create a catalog of vetted safety, security, and trust tools to aid developers in ensuring responsible AI development.
Members of the alliance have pledged to work on open foundation models, focusing on highly capable multilingual, multi-modal, and science models. These efforts are aimed at addressing global challenges such as climate change and unequal education opportunities. However, the alliance is yet to establish a governing board and technical oversight committee and is currently leaning on the Partnership on AI's Model Deployment Guidance to manage the use of open models.
While the AI Alliance promotes the benefits of open source AI, it faces criticism. Some argue that simply releasing a model with acceptable use policies is not enough to prevent malicious use. There are also concerns that some open source models, particularly those from Meta, are not sufficiently open due to license restrictions.
In contrast, TechNet, a group of senior American tech executives, has launched a campaign emphasizing the positive impacts of AI on American lives. They argue that the AI conversation should focus more on practical applications rather than hypothetical risks.
Key figures in the AI industry have voiced their support for the alliance. CJ Desai, president and COO of ServiceNow, emphasized the need for collaborative approaches in AI, while IBM Chairman and CEO Arvind Krishna highlighted the importance of open innovation for ensuring safety and scientific rigor. Nick Clegg of Meta also stressed the significance of open research collaboration in advancing AI development.