A new coalition of over 50 leading companies, academic institutions, and non-profits has come together to create the AI Alliance, a global community focused on fostering open innovation in artificial intelligence. Launched by IBM and Meta as co-founders, the AI Alliance brings together major players across the AI landscape with the aim of accelerating AI progress through collaboration while ensuring developments adhere to principles of trust, transparency, diversity, and responsibility.
The Alliance boasts a diverse and prestigious membership, including:
- Leading universities and research agencies driving AI education and research.
- Innovators in AI hardware and infrastructure, such as GPU manufacturers and cloud platform providers.
- Champions of software frameworks including PyTorch, Kubernetes, Transformers and Diffusers.
- The creators of the tooling driving AI benchmarking, trust and validation metrics and best practices such as MLPerf, Hugging Face and LangChain
- Creators of widely used open models like Llama2, Stable Diffusion, StarCoder, Bloom, and many others.
As outlined on their website, the AI Alliance has a multifaceted mission centered on enabling open AI ecosystems. This includes goals such as building open source tools for AI development, creating catalogs of vetted resources for evaluating AI safety, advancing multilingual AI models to serve more communities, boosting hardware contributions for AI training, engaging students and researchers through educational initiatives, and advocating for policies supporting open and responsible AI. The key objectives include:
- Promoting Open Technologies: The Alliance focuses on building support for open technologies across software, models, and tools. It encourages experimentation and adoption among developers and scientists, advocating for open innovation at various organizational and societal levels.
- Safeguarding AI: By developing benchmarks, tools, and methodologies, the Alliance ensures the responsible use of AI. It emphasizes the creation of a vetted catalog of safety, security, and trust tools, aligning with meaningful benchmarks and evaluation standards.
- Enhancing Foundation Models: The Alliance is committed to advancing an ecosystem of diverse open foundation models, targeting challenges in areas like climate and health. This involves the collaborative development and promotion of open-source tools for model training, tuning, and inference.
- Advancing AI Hardware Ecosystem: Collaborative efforts are aimed at benchmarking and optimizing AI workloads to support a diverse range of hardware, focusing on scalability and advanced modeling for power, energy, and carbon footprint.
- Global AI Skill Building: Engaging the academic community is key. The Alliance supports research projects and develops educational content to close the AI skill gap, informing public discourse and policymakers on AI benefits, risks, and regulatory approaches.
The group consists of over 50 launch partners spanning industry giants, research pioneers, and innovators in AI hardware, software, and applications. Founding members include AMD, Intel, Linux Foundation, Sony, Oracle, IBM, Meta, NASA, Stability AI, AI21 Labs, Hugging Face, and many top universities such as Cornell University, Berkeley, UT Austin, Yale, and more.
While an impressive cast, notably absent from the list are the leading AI research labs and model providers such as OpenAI, Anthropic, Google, Microsoft, NVIDIA, Inflection, Cohere, Amazon and others.
In quotes included with the Alliance’s launch announcement, IBM CEO Arvind Krishna emphasized the importance of open collaboration in AI, stating “progress we continue to witness in AI is a testament to open innovation and collaboration across communities of creators, scientists, academics and business leaders.” Meanwhile, Meta President of Global Affairs Nick Clegg highlighted the goal of making AI innovations accessible beyond large tech companies, saying “We believe it’s better when AI is developed openly — more people can access the benefits, build innovative products and work on safety.”
The AI Alliance will operate through member-driven working groups focused on key AI areas has already begun convening working groups and laying plans to start on projects supporting its goals. A governing board and technical oversight committee will guide the projects, establishing standards and guidelines. Collaborations with existing governmental, non-profit, and civil society initiatives will further enhance the Alliance's impact.
Open innovation in AI democratizes access to advancements, harnessing global talent and ensuring accountability and transparency. This approach is crucial for robust testing, validation, and broad-based community approaches in AI.
While still coming together in its early stages, the high profile backers and commitment to advancing open AI signals this may become an influential force in shaping the future AI landscape to be more collaborative, transparent, and ethically-grounded. The coming months and years will reveal how impactful the AI Alliance’s contributions ultimately prove to be.