For all of the challenges and problems that generative AI (GenAI) thrust upon American society, those closest to the rapidly evolving technology stop short of the doom and gloom scenarios, and instead urge the patching of an imperfect system.
During a “fireside chat” with Gov. Gavin Newsom at the Joint California Summit on Generative AI last week, academic experts shared what they saw as the true threats of the technology and what can be done to maintain the state’s lead in the space. “How the hell do we stay on top of this?” Newsom even asked at one point.
Fei-Fei Li, a co-director of the Stanford Human-Centered AI Institute (Stanford HAI), said that while the human extinction-level scenarios are overplayed “sci-fi,” the real threats posed by the technology center on things like labor market shifts, democratic unrest sown by misinformation, bias and privacy issues.
“I think we have a very strong stance, especially coming from Stanford HAI, that we believe AI has to augment people; it cannot replace people. It’s not a technology that should take away our collective prosperity as well as individual dignity,” Li said.
“I think these are much more real problems that a lot of our multi-stakeholders in this audience are really working so hard on,” Li added.
During the roughly hourlong conversation, the panelists touched on the importance of democratizing the technology, while adjusting how it fits into all communities and the public sector.
Jennifer Chayes, dean of the UC Berkeley College of Computing, Data Science and Society, agreed with Li, noting that one of the major issues in the emergence of AI has been access limitations — much like those created by Internet deserts.
“I really believe GenAI can be a great democratizer if we make sure that we empower everyone to be able to use GenAI to solve their problems, the problems of their communities, and in so doing we will find where it’s going off track because there must always be a human in the loop …,” Chayes said.
Li and Chayes both called for a “moonshot mentality” toward the technology — combining the collective might of academia, government, nonprofits and industry. “The truth is that [the] public sector is one of the driving forces of this country’s innovation,” Li said, pointing to the Internet, GPS and biotech advances as examples of past innovation.
Chayes said the technology could be made better with fairer outcomes for constituents by separating — and opening — the GenAI models used in government from the ones developed in the private sector. The wealth of government data and open and transparent generative systems could be a boon in areas like social services, health care and more.
“I believe that the solution to a lot of the risks is putting it out there, building things as communities and testing the heck out of it as we’re doing so,” Chayes said.
Educating at scale was one of the critical steps Chayes pointed to with regard to maintaining the state’s lead in the space. Berkeley, for its part, has seen a nearly 50 percent increase in applications for computer science, data science and statistics degrees for the upcoming year, Chayes said.
Newsom pointed to the advantage that the state has in attracting global talent: “We believe in inclusivity, in equity, we believe in opportunity, and we want folks from around the globe to feel seen and heard, and that’s why this state has dominated for decades, because we are a majority-minority state. Twenty-seven percent of the state is foreign-born. …”
The governor also called out the state and federal legislative knowledge gaps. California currently has some 20 bills related to AI moving through the Legislature, Newsom noted, each with its intent and ramifications and written by lawmakers with varying levels of understanding.
“We don’t want to screw this up, but we don’t want it to screw us up,” he said.
