[News Terms] United Against 'OpenAI', the 'AI Alliance'

On the 5th, a new coalition was launched to challenge OpenAI, the creator of ChatGPT, which is leading the generative artificial intelligence (AI) market. This is the 'AI Alliance,' a group consisting of more than 50 AI-related companies and institutions. Participants include Meta, the parent company of Facebook, IBM, Intel, AMD, Oracle, Dell, Sony, SoftBank, and semiconductor, cloud, and device manufacturers from the United States, Japan, and Europe. National agencies such as NASA and the National Science Foundation (NSF), as well as universities including Harvard, Yale, Cornell, Berkeley, Boston, Dartmouth, Notre Dame, University College London in the UK, Hebrew University of Jerusalem in Israel, and Keio and the University of Tokyo in Japan, are also part of this alliance.

According to foreign media reports, the core goals of the AI Alliance are to break monopolies and promote openness and transparency. They emphasize that the future of the AI industry will fundamentally be built on collaboration and open innovation regarding AI models. To counter the closed approach of OpenAI and Microsoft (MS), who do not disclose their development source codes externally, the alliance plans to share their AI achievements and accelerate AI development as a form of collective intelligence open to anyone. The AI Alliance declared that it will "engage in activities to ensure scientific rigor, trust, safety, security, diversity, and economic competitiveness."

A distinctive feature of the AI Alliance is that big tech companies and academia provide large language models (LLMs) as open source. For example, Meta publicly released its own LLM, 'Llama 2,' in July, making all related technologies available for commercial use. Since then, over 100,000 universities and startups worldwide have downloaded Llama and developed high-performance AI.

In contrast, the closed AI camp, including OpenAI, MS, and Google, do not disclose data or software externally nor share them with other companies. They argue that careless sharing and openness of AI technology could lead to misuse or pose risks to humanity.

The AI Alliance aims first to distribute benchmarks, tools, resources, and open models to support responsible AI development and global-scale AI usage. It is also focusing on AI-related regulations and safety, with plans to soon release tools for AI safety and model verification.

This content was produced with the assistance of AI translation services.