But many of the world’s biggest AI players, including Microsoft, Microsoft-backed OpenAI, Google, and Amazon Web Services (AWS ), all long-standing defenders of closed source AI models and the proprietary, licensed approach, are absent from the alliance.
Intriguingly, however, Partnership on AI (PAI), a nonprofit coalition announced in 2016 by founding members Amazon, Facebook, Google, DeepMind, and Microsoft, has also joined the alliance.
The coalition, which is committed to the responsible use of AI, said it will continue to work with its 90+ partners and the alliance members to ensure that “open science and innovation contribute to safe and responsible AI.”
While open source AI has been riding the wave of growth, Microsoft has only sluggishly embraced the movement, partnering, for instance, with Meta’s Llama 2 and Databricks to bring open-source AI models to its Azure cloud platform. Google, AWS and OpenAI have moved slowly with their open source AI projects.
On the other hand, Meta’s entry in the AI arms race, largely dominated by heavy hitters like OpenAI, Google, and Microsoft, was marked by its commitment to market open source models, notably with the release of Llama 2, one of the most used foundation models.
IBM also recently released an open source foundation AI model, in collaboration with NASA, also a member of the alliance, designed to help scientists analyze satellite imagery.
“We believe it’s better when AI is developed openly – more people can access the benefits, build innovative products and work on safety,” said Nick Clegg, president, global affairs, Meta. “The AI Alliance brings together researchers, developers and companies to share tools and knowledge that can help us all make progress whether models are shared openly or not.”
IBM and Meta bring different capabilities to their partnership headlining the alliance, said Brian Jackson, principal research director at Info-Tech Research Group.
Jackson noted that IBM brings its credibility in professional services and experience in taking open source software – most notably with its 2019 Red Hat acquisition – and translating it into a solutions and services business model. Meta, in contrast, brings its AI talent and leading researchers to the alliance.
This focus on open source through the alliance, however, will likely be more advantageous to Meta than to IBM, he said.
“For Meta, its open-source models are already available on Microsoft’s Azure platform. Its models will find enterprise users so long as there are professionals willing to support them. That takes some wind out of the sails of IBM, who can’t claim any sort of competitive advantage by hosting Meta’s models. At best, it can claim that it has more support services built up around them.”
Chief product officer (CPO) of information management giant OpenText, Muhi Majzoub, acknowledged that IBM and Meta coming together to form the AI Alliance leaves a lot of question marks for the enterprise.
“Any CIO will tell you, ‘I’m not putting my data in a publicly trained, open source LLM’. Business leaders are looking for AI solutions where their company’s data is secure and protected.”
But while AI researchers worry that cybercriminals will download and use these open-source models to further illicit business models, that’s not necessarily a risk for organizations using open source models from the AI alliance today, contended Jackson.
He added, “But in the future, if governments apply more restrictions to the release of LLMs, then it could become an issue.”
Notable presences in the newly formed Alliance are Hugging Face, Oracle, ServiceNow, Dell, Linux Foundation, Stability AI, Red Hat, and several startups, leading universities, governments and more.
Interestingly, the Alliance also includes chip giants including AMD, Intel, and Cerebras, who, since the AI boom, have been steamrolled by competitor Nvidia, currently occupying a near monopoly in AI-enabled GPUs and selling to the biggest cloud providers like AWS, Azure and Google.
AMD’s chief executive Lisa Su said, “The history of our industry highlights how open, standards-based development leveraging the capabilities of the entire industry both accelerate innovation and ensure technology advances have the largest positive impact.”
What will the Alliance do?
IBM said that the AI alliance will begin with the following projects:
- Develop and deploy benchmarks and evaluation resources that enable the responsible development and use of AI systems
- Responsibly advance the ecosystem of open foundation models with diverse modalities to address society-wide challenges
- Foster a vibrant AI hardware accelerator ecosystem
- Engage the academic community to support global AI skills building and exploratory research
- Develop educational resources to inform the public discourse and policymakers on AI regulation
- Host events to explore AI use cases and showcase how Alliance members are using open technology in AI responsibly.
The AI alliance, IBM said, will begin its work with the formation of member-driven working groups across the areas listed above. It will also establish a governing board and technical oversight committee tasked to advance the key project areas and establish project standards and guidelines.
Further, the alliance plans to partner with existing AI initiatives from governments, non-profit and civil society organizations.