Major players in the technology industry, including IBM and Meta, have joined forces to establish the AI Alliance, a collaborative initiative aimed at advancing the development of artificial intelligence (AI) models. This alliance brings together not only tech giants but also government entities, research organizations, and startups, all committed to fostering open innovation and open science in AI, as stated in a press release by IBM.
The surge in AI development, particularly since the release of OpenAI’s ChatGPT, has prompted a competitive race among technology companies to introduce their own AI models capable of handling text, visual, and audio content through machine learning. While the road to achieving artificial general intelligence (AGI) remains distant, a few key industry leaders have emerged, leaving legacy technology companies like Dell, Sony, Oracle, and chipmakers such as AMD and Intel trailing behind. Through the AI Alliance, these companies aspire to collaborate and catch up with pioneers like Microsoft-backed OpenAI and Google.
Emphasizing the importance of open-sourced AI, the AI Alliance members underscored the need to prioritize safety, diversity, economic opportunity, and benefits for all. Dario Gil, IBM’s Senior Vice President, expressed concern about the narrow focus of certain institutions in AI over the past year without specifying names.
The AI Alliance comprises research institutes such as A*STAR, CERN, NASA, and numerous global universities. Its members are the creators of widely used open-source models such as Llama2, Stable Diffusion, StarCoder, Bloom, and others. The alliance aims to meet the demand for responsible AI development, involving stakeholders in research, development, deployment, and governance.
The mission of the AI Alliance includes creating a catalog of safety, security, and trust tools for global AI deployment. It seeks to collaborate with developers to implement these tools during model and application development. Additionally, the alliance plans to develop multi-modal and multi-lingual models addressing current societal challenges like climate and education. Collaboration with research universities aims to enhance AI skills globally, fostering greater participation in AI modeling projects.
The alliance members aspire to generate educational content for users and policymakers, shedding light on the benefits, risks, and regulation of AI. They plan to engage in events to showcase the advantages of open-sourced AI to a broader audience.
While the AI Alliance has initiated the formation of working groups for various areas of focus, it is still in the process of establishing a governing board and a technical oversight committee. Nick Clegg, President of Global Affairs at Meta, stressed the importance of open development in AI, stating that the AI Alliance aims to facilitate collaboration among researchers, developers, and companies to advance AI’s state-of-the-art and promote responsible practices.