Meta is collaborating with IBM to launch the AI Alliance, a new industry body focused on supporting “open innovation” and “open science” in AI. This move aims to educate the public about AI issues and develop open AI resources that equally and responsibly meet the needs of business and society. The AI Alliance, which includes members like the Partnership on AI, will establish working groups, a governing board, and a technical oversight committee to advance areas such as AI trust and validation metrics, AI training infrastructure, and open source AI models and frameworks.

The Alliance’s approach, although seemingly overlapping with the objectives of its members’ individual efforts, is intended to be complementary and additive. It emphasizes collaboration and information sharing to mitigate specific risks and innovate more inclusively and rapidly. The AI Alliance takes a jab at companies with a closed, proprietary vision for AI, highlighting the importance of an open platform for AI innovation.

Meta’s open source AI approach has faced criticism from companies like Google, OpenAI, and Microsoft, but it continues to take calculated risks, releasing models like Llama despite potential misuse. The Alliance’s co-founder, IBM, likely seeks more exposure for its generative AI platform amidst competition from Microsoft, OpenAI, and Google.

The AI Alliance’s early membership notably omits certain high-profile organizations and companies, reflecting philosophical differences and strategic choices. The Alliance, currently comprising about 45 organizations, aims to foster an open community and enable responsible AI innovation, emphasizing safety, security, diversity, and economic competitiveness.

Despite the absence of major AI industry players and a clear path to success, the AI Alliance’s diverse membership across sectors like healthcare, silicon, software-as-a-service, and academia positions it uniquely at the intersection of AI and enterprise. Its success, however, remains uncertain amidst competing interests and skepticism about the sincerity of its founding members, Meta and IBM.