IBM and Meta have come together to launch the new AI Alliance, a collaboration involving over 50 prominent organizations from around the world. With industry leaders, startups, academic institutions, and government bodies joining forces, the AI Alliance aims to support open innovation and open science in AI. This article will explore the objectives and key focus areas of the AI Alliance, highlighting the importance of responsible innovation, diversity, and economic competitiveness in the AI landscape.
A Coincidental Timing
Sriram Raghavan, vice president of IBM Research, dismisses the notion that the timing of this launch is influenced by recent events within the AI community. Though it may appear prescient given the drama at OpenAI and the ongoing negotiations of the EU AI Act, Raghavan insists that discussions for the AI Alliance began in the summer. The motivation behind this initiative is rooted in the need for a more nuanced and comprehensive discussion around AI. Raghavan highlights the shift towards closed, proprietary AI development and concerns about stifling innovation due to debates surrounding AI risk. The intention of the AI Alliance is to fill the gap by focusing on open innovation and developing solutions that address the industry’s needs.
Expanding the AI Narrative
Raghavan emphasizes the need to broaden the AI narrative beyond discussions of risky models and malicious intent. While acknowledging the risks associated with unrestricted access to AI models, he believes that building a small group of proprietary institutions and introducing excessive regulation is not the solution. Instead, Raghavan argues for the development of safe and responsible AI that can be done openly. By coming together, the AI community can define benchmarks for safe AI behavior and ensure that models are developed and utilized responsibly.
Nick Clegg, president of global affairs at Meta, supports the idea of open AI development for its wide accessibility, innovative potential, and focus on safety. The AI Alliance brings together researchers, developers, and companies to share tools and knowledge, regardless of whether models are shared openly or not. The goal is to advance the state-of-the-art in AI and promote responsible practices within the industry as a whole.
The AI Alliance operates with a project-based and flexible approach, concentrating on six main areas:
1. Developing Benchmarks and Evaluation Standards
One of the primary objectives of the AI Alliance is to create benchmarks and evaluation standards that enable responsible development and use of AI systems on a global scale. Additionally, the alliance aims to build a catalog of trusted safety, security, and trust tools that can support developers and researchers in building and deploying AI models and applications.
2. Advancing Open Foundation Models
To foster a diverse ecosystem of open foundation models, the AI Alliance focuses on developing highly capable multilingual, multi-modal, and science models. By addressing challenges in areas such as climate and education, these models can contribute to societal advancements.
3. Boosting AI Hardware Accelerator Ecosystem
The AI Alliance recognizes the importance of an efficient hardware accelerator ecosystem to facilitate advancements in AI technology. By promoting the adoption of essential enabling software technologies, the alliance aims to drive contributions and growth in this sector.
4. Supporting Global AI Skills Building
Engaging with the academic community, the AI Alliance encourages research and education in AI. By supporting researchers and students in AI model and tool research projects, the alliance contributes to the development of AI skills globally.
5. Informing the Public Discourse
To ensure a well-informed public discourse and effective policymaking, the AI Alliance develops educational content and resources. These efforts aim to inform individuals and policymakers about the benefits, risks, and solutions related to AI, as well as discuss the need for precise regulation.
6. Encouraging Open AI Development
The AI Alliance actively encourages open development of AI in safe and beneficial ways. The alliance hosts events to explore various AI use cases and showcase how its members utilize open technology responsibly and for societal good.
The AI Alliance represents a significant partnership of influential organizations dedicated to fostering responsible and open innovation in AI. By pooling resources, sharing knowledge, and prioritizing safety and diversity, this global collaboration aims to address the challenges and risks associated with AI while promoting scientific rigor and economic competitiveness. As the AI landscape continues to evolve, the AI Alliance stands as a testament to the commitment of its members in shaping the future of AI for the benefit of society.