White House to host event showcasing AI’s potential

On Thursday morning the White House will convene agencies, industry and academia to discuss the federal government’s research efforts for practical artificial intelligence applications in multiple public sector contexts, including weather models, education efficiency, materials development for semiconductors and energy grid resilience.

Arati Prabhakar, the director for the White House’s Office of Science and Technology Policy, briefed reporters during a Wednesday press call on what to expect from today’s AI Aspirations: R&D for Public Missions discussion. The conference will highlight the Biden administration’s plans to leverage AI and machine learning systems to benefit pressing public service needs.

“AI can help us deliver critical services to any American right when they need them most,” Prabhakar said. 

The event will highlight AI research and development efforts for application in seven different sectors: drug discovery; individualized education; meteorology forecasts; electrical grid security; materials development for semiconductors; transportation infrastructure projects; and improving government services. 

Prabhakar said that this effort will be spearheaded by experts in the federal government as well as in labs, academia and the private sector. The conference will feature seven individuals from the federal government but include external participants.

Iambic Therapeutics, Microsoft, and GE are among the “huge range” of companies from different sectors participating with OSTP, Prabhakar said. 

“This is a conference about projecting these big visions, and each of them is going to be a series of steps to get from here to the ultimate vision, the things that really do change Americans’ lives,” she said. 

A key feature of the seven individual AI applications is the diversity of AI models and the data they will require to function properly. Prabhakar noted that some of the application models will learn from preexisting language, while others will be trained on biological data from previous clinical trials or data from barometric sensors and radar systems. She said that researching how to bring beneficial AI capabilities across multiple sectors will help inform ongoing efforts to standardize testing benchmarks for various AI and machine learning models.

“We'll learn a lot from all of these different kinds of AI models,” she said.

Timelines for completion in each application vary, given the sensitive nature of the use cases –– especially drug development –– and are envisioned as multiyear projects that will feature a number of research and development efforts.

Building upon research already conducted by the U.S. national lab network will feature in the AI Aspirations work.

“To do this work well is going to require all the marvels of AI, and it's going to continue to require the laboratory research that's been the foundation of so many of these different disciplines,” she said.