For the U.S. to maintain a competitive edge in the global artificial intelligence race, Sen. Maria Cantwell (D-WA) says there are lessons to take from post-World War II.
Speaking at a Tech Alliance event in Seattle on Friday, the former tech executive and longtime senator drew comparisons to the G.I. Bill that helped train World War II veterans as they entered the workforce.
“Instead of a G.I. Bill, we need an AI Bill for education,” said Cantwell, who chairs the Senate Commerce Committee.
Such a bill would help people working in various sectors gain skills related to AI, Cantwell said.
The former RealNetworks executive specifically called out prompt engineering, or the practice of crafting inputs for generative AI systems to produce desired outputs. She wants to find ways to accelerate skills such as prompt engineering for the next generation of the workforce.
Cantwell drew another parallel to World War II, noting the women codebreakers who helped provide critical intelligence information.
“They were science and math teachers from West Virginia … they married both the science and a little bit of the intuitiveness together, to then figure out this code,” she said.
The senator spoke on a keynote panel with Ali Farhadi, CEO of the Allen Institute for AI (AI2) in Seattle, moderated by Dave Cotter of D3 Advisors, a startup founder and veteran of companies including Leafly and Amazon.
Farhadi pointed to the importance of multidisciplinary centers and tapping the expertise of people from various domains as it relates to AI development.
“If you want to get state of the art AI and extend it to a new domain — cell biology, neurology, atmospheric sciences, agriculture — you actually need to change so much of the inner workings of these systems,” he said. “That means that you need to have expertise both on the science side, but also on the AI side. And that is both a unique opportunity, but also a unique challenge.”
Farhadi said AI education is “radically different” from teaching traditional coding and software. He noted the huge computing capacities needed to “play with one of these systems at a meaningful scale,” and said resources provided by government are crucial.
Farhadi also said it’s key that a broad swath of Americans get access to AI education, particularly given the problems of bias in AI.
“We want people who are designing or writing the code or collecting the data to be of diverse backgrounds,” he said. “We need to find a way to educate people of diverse backgrounds about these skills. That’s the biggest challenge we’re going to face, on who to educate with AI.”
