Fei-Fei Li, also known as the “godmother of AI” has been articulating to policymakers the urgent need for investment in AI research at the national level. Currently working as a professor at Stanford University, she is acutely aware of the fact that the public sector is falling behind in its ability to perform progressive research due to a lack of resources including data, GPUs, computing power, and the inability to enlist the right people.
The State of Affairs
Li has been advocating for investments into a national warehouse that would serve as the host to computing power and datasets for the public sector. She is looking to make steps forward in the priority put into this field by the United States government as companies like Meta, Google, and Microsoft continue funneling billions of dollars into AI research. “The public sector is now significantly lagging in resources and talent compared to that of industry,” said Li. “This will have profound consequences because industry is focused on developing technology that is profit-driven, whereas public-sector AI goals are focused on creating public goods.”
The Resource Gap Widens
The public sector and leading universities are unable to keep up with this level of investment, and the results are beginning to show. Vital resources such as GPUs, a type of specialized computer chip, are severely lacking among researchers outside of these major companies. As Meta aims to procure 350,000 GPUs in the near future, Stanford’s Natural Language Processing Group has only 68 GPUs for all of its work
Physical resources are not the only thing universities and government initiatives cannot obtain. Many of the country’s would-be computer science professors are being drawn away from campuses to high-rises as companies are offering high salaries and the chance to work on the frontier of AI technological advancement. Two decades ago only 21 percent of people with PhDs in AI would end up in the private industry. Today, there is a steep contrast to that level with around 70 percent of PhD’s working in private industry.
Consequences of the Gap
Historically, large tech companies have run their AI research labs in a manner that resembles universities. Scientists and researchers would often decide what project was a priority to advance the state of research. They were mostly isolated from teams focused on other aspects such as building products or generating revenue, creating an optimal environment for organic findings judged by notable breakthroughs and papers published similar to the metrics used for their university peers.
However, as we move into an increasingly competitive market for generative AI products, that could all change. For example, Google announced that it would take steps to merge its more research-based teams with its product development teams. This department known as Google DeepMind, has already started to take more advantage of its own AI discoveries by sharing research papers only after the lab work had been turned into products. Meta is also adopting a similar approach to combine research and development.
The Create AI Act
Li is just one of a growing chorus of members from universities, public organizations, and the government looking to enhance AI research capabilities outside of private hands. Another leader of this movement to democratize AI is U.S. Representative Anna G. Eshoo who introduced the Creating Resources for Every American To Experiment with Artificial Intelligence Act of 2023 (CREATE AI Act). The goal of the CREATE AI Act is to establish the National Artificial Intelligence Research Resource (NAIRR) as a shared national research infrastructure that provides AI researchers and students from diverse backgrounds with greater access to complex resources, data, and tools needed to develop safe and trustworthy artificial intelligence.
Benefits of the CREATE AI ACT
- Access to Resources: The act would work to allocate funding on a federal level to establish AI research institutes across the country. Serving as AI hubs, they would be home to research, innovation, and education providing AI resources to researchers and students to foster advancements in the field.
- Collaboration and Networking: Research institutes would provide the opportunity for greater communication among researchers, students, industry professionals, and policymakers. The ultimate acceleration in innovation and adoption of AI would be a key result of collaboration on research projects, knowledge sharing, and an exchange of best practices.
- Ethical Considerations: Studies on the ethical implications of AI technologies will be used to develop guidelines and frameworks for responsible AI use. The promotion of awareness and accountability will mitigate potential risks associated with AI deployment while maximizing its societal benefits.
The CREATE Act has the potential to significantly enhance the research and educational landscape of artificial intelligence in the United States. It can empower researchers and students to explore new frontiers in AI innovation without the need for funding, more importantly, product-based funding, from large corporations looking to make a profit.
