Fifty leading companies have decided to join forces to support solutions artificial intelligence that use the open source approach. The new initiative is called AI Alliance and is focused on promoting an open community, building on the ability of developers and researchers to accelerate responsible innovation in the field of artificial intelligence. At the same time, AI Alliance wants to ensure scientific rigor, trust, safety, diversity and economic competitiveness.
The objective of the newly formed “alliance” is in fact to bring together developers, scientists, academic institutions, companies and other leaders capable of innovating, pooling resources and knowledge to address the problems and challenges that intelligence solutions artificial they propose to us.
Among the names of the participants in the AI Alliance, those of IBM, Meta, AMD, Intel, Cerebras, Hugging Face, Oracle, Red Hat, Linux Foundation and Dell Technologies certainly stand out.
AI Alliance: For an AI to be safe, it must be open
Examining the content of the statements released by the various members of the newly formed AI Alliance, an interesting fact emerges: the idea is to build Safe AI comparing each other and defining, by mutual agreement, reference parameters.
In other words, according to the promoters of the initiative, it makes no sense build models and then cage them within a severe regulatory enclosure. Instead, we need to promote the diametrically opposite path, that is develop openly so that anyone can access the benefits, build innovative products and work on security.
As explained on the project website, open and transparent innovation is essential to provide a wide range of researchers, developers, software houses with the information and tools necessary to exploit these advances in order to give maximum priority to safety, diversity, opportunities economic and beneficial for all.
The main objectives of the AI Alliance
On the basis of what was highlighted by the founders of AI Alliance, the agreement has as its first aim precisely that of development and implementation reference parameters and evaluation standards. These tools will enable the responsible development and use of AI systems on a global scale.
AI Alliance also aims to support the promotion and enablement of the tools thus identified, among the developer community in order to create models and applications.
We also talk about the use of foundation modelsincluding multilingual, multimodal and scientific ones, which can help address societal challenges in climate, education and more.
The work done within the AI Alliance can also foster a vibrant ecosystem of AI hardware acceleratorssupport the adoption of enabling software technologies, involve the academic community, encourage research activities.
In May 2023, it was said that open source models for artificial intelligence would surpass those of OpenAI and Google. Now the market is full of open solutions: AI Alliance aims to take on the delicate role of “conductor of the orchestra”.
Opening image credit: iStock.com/Black_Kira