Newly-founded AI Alliance promotes collaboration on open and safe AI technology

Alfonso Maruccia

Posts: 1,025   +301
Staff
In context: ChatGPT, the gem in OpenAI's extensive array of generative AI services, was predominantly developed using community-oriented, open-source technologies like TensorFlow and PyTorch. The existence of the ChatGPT business owes much to these FOSS tools, even as the company has shifted its primary focus towards transforming its machine learning algorithms into a profitable venture.

While major technology companies like OpenAI, Microsoft, and Google have adopted a predominantly closed, private approach to AI technology, a significant portion of the modern technology industry appears inclined to take a different path. IBM and Meta/Facebook have recently introduced the AI Alliance, an international community dedicated to harnessing AI as a force for good rather than merely for the benefit of shareholders.

According to the official announcement, the newly-launched AI Alliance boasts over 50 founding members, including leading chipmakers such as AMD and Intel, prominent technology companies like Oracle, Dell, Red Hat, and Sony Group, as well as various AI startups like Stability AI, Hugging Face, and Silo AI. Additionally, the Alliance encompasses numerous research organizations and world-class universities across the globe.

While the AI Alliance does not explicitly mention companies like OpenAI or Google, which may prioritize AI as a business endeavor, it clearly advocates for an alternative approach. The organization is uniting to champion open innovation and open science in AI, as stated in the official press release. Its goal is to shape the development of generative algorithms and transform them into tangible, beneficial solutions that address the needs and complexities of our societies.

The AI Alliance aims to unite developers, scientists, academic institutions, companies, and "other innovators," pooling resources and knowledge to address AI safety concerns. The Alliance's "action-oriented" projects will include the development of benchmarks and evaluation standards for responsible AI systems and the creation of open foundation AI models to tackle society-wide challenges in climate, education, and beyond.

Additionally, the AI Alliance will strive to foster a "vibrant AI hardware accelerator" ecosystem, currently dominated by powerful GPU accelerators from Nvidia. Its objectives also encompass supporting global AI skill-building, creating educational content, and providing other resources to contribute to the public discourse on AI. The organization plans to launch initiatives promoting open AI development and will host events to explore AI use cases.

Before embarking on its many ambitious goals for open AI development, the AI Alliance will first establish its governing board, a technical oversight committee, and standards and guidelines for the overall project. The organization also intends to collaborate with government institutions, non-profit organizations, and civil society groups already contributing to the promotion of responsible AI advancements.

Permalink to story.

 
My fear is that a novel, partially-government funded, multi-functional, self replicating, chimera AI will (accidentally or purposefully) escape the lab and wreak havoc on the world.
 
Back