Access to AI Resources Is Not Equal
Simply put, the scientific computing resources needed to tap into the power of AI remain unavailable to those who need it most. The system may end up shutting out the next generation of Albert Einsteins.
Gone are the days when patent clerks, garage tinkerers and dreamers who lack a spare $50,000 lying around can make the next great innovation — unless we do something about it. Scientific research must be democratized by leveling the playing field and giving all researchers access to powerful scientific computing.
NVIDIA does offer an entry-level “hobbyist” chip, but it’s only powerful enough to use existing models, not to help scientists develop their own.
We miss out on generational brilliance when we limit innovation to those with fancy titles and financial backers, because the best ideas come from everywhere.
DISCOVER: Artificial intelligence pervades early ARPA-H medical moon-shot projects.
How To Democratize AI’s Use in Scientific Research
For years, the National Science Foundation has funded AI research supporting the development of technologies that are instrumental for the scientific community. However, like any source of federal funding, the amount of money available is limited and highly competitive.
A significant expansion in funding is needed to make a real difference in the affordability and accessibility of AI technology for smaller labs and independent researchers. This seems unlikely to happen. Rather, there is more potential for public-private partnerships between institutions with computing resources and researchers who need them.
Institutions with the proper infrastructure can loan their power to researchers seeking it, granting them access to high-performance computing at a fraction of the price it would cost for them to purchase and maintain their own dedicated hardware. This is the kind of collaboration the scientific field is built on, and it will lead to a brighter future, where AI supercharges research.
