Biden’s Big Push to Democratize AI Arrives. What Does It Mean for Your Company?

Though artificial intelligence applications now reach into many areas of life and work, thanks to innovations like ChatGPT, work on next-generation AI tech remains tricky and expensive. Researchers doing the really clever stuff often require access to big, powerful computer systems that come with steep price tags. That’s what the now-live National Artificial Intelligence Research Resource, NAIRR, program aims to tackle. The sheer size of NAIRR underlines how serious this new AI push really is: There will be some $800 million in funding available over three years.

The program is backed by the National Science Foundation and 10 other agencies in a partnership. It’s all about providing access to AI models, raw computer power, datasets that could be used for training, software and even training for people researching AI. Even more interesting, companies like Nvidia, the graphics chip making company whose chips are now sought for AI processing purposes, are also chipping in with funding–$30 million from Nvidia alone. Tech giants Microsoft, Google, and Meta are involved, as are other industry partners. AI industry leader OpenAI is also contributing resources. 

The director of the NSF’s office of advanced cyberinfrastructure explained that the goal was to boost AI research at “smaller institutions, rural institutions,” and those representing underrepresented populations. To start with, only a small number of proposals for NAIRR money will be accepted, but “hundreds” more opportunities will open up in the spring, the agency said.

The whole program is split into four separate segments. The NAIRR Open piece is about enabling access to “diverse AI resources” to boost open AI research. For context here, Meta just revealed it was pushing to develop real artificial general intelligence, and that it would open-source some of its research. There’s also a secure segment of NAIRR for research focusing on privacy and security. The software segment is aimed at boosting interoperability of different AI tools, and the classroom segment is aimed at AI uses in education and training.

The NAIRR program is unlikely to impact your company directly, unless AI research happens inside your engineering team and the research topic is compatible with what NAIRR systems are available for–such as OpenAI’s donating access to models to help boost AI safety. NAIRR does underline the fact that AI is going to be ever more prevalent in society and business, possibly impacting the workplace and replacing workers. 

The program may also offer scope for small, agile businesses–particularly those working in remote communities–to work with AI researchers in institutions. Either by providing services that help with AI research, or in business partnerships where the goal is to test out real-world enterprise AI applications.

And if this entire endeavor sounds somewhat ephemeral and political, with generic-sounding goals like bolstering the U.S.’s “global competitiveness” in the AI field, NSF Director Sethuraman Panchanathan says that it’s worth remembering the origins of the internet and the web. 

The internet’s complex global infrastructure of servers and network of connecting cables sprang from research projects at the federal Advanced Research Projects Agency from the late 1960s on. And the World Wide Web–the system of digital communications sent over the internet that powers, for example, this website you’re reading–came from innovations at Europe’s Nuclear Research facility, CERN. 

NAIRR could be a similar government body that brings AI into the mainstream in the same way ARPA’s network became the internet, and jumping on board early could be a good thing for your business in the long term.