As California expands safeguards for artificial intelligence and social media, USC students and professors have mixed feelings about regulation.
Just this past week, Gov. Gavin Newsom signed legislation to ban addictive social media features, such as autoplay or user-generated algorithms, regulate companion chatbots and manage the use of student data in AI systems.
Calls for regulation were sparked by researchers concerned AI is advancing too quickly.
Former Anthropic and OpenAI researcher Jacob Coxon posted on X that he was resigning because he was concerned about the threat AI poses.
“[OpenAI and Anthropic] will soon be superhuman systems that can hack anything, revolutionize any field overnight, and acquire real power and resources,” Coxon wrote.
His warning has sparked debate over unchecked regulation of AI systems.
Stephen Chih-Yang Lu, a professor of Industrial and Systems Engineering, Aerospace and Mechanical Engineering and Computer Science at USC, said that it’s “too late” for the government to get involved.
“Everybody wants to be ahead,” Lu said. “I’m not sure about this regulation. Even if the government starts imposing it, how is that going to turn out? And in a way, in my view, the genie is already out of the box.”
Jaspreet Ranjit, a PhD student majoring in computer science, said she believes federal involvement in regulating artificial intelligence is needed.
“Going off of what Newsom has mentioned, California can be a really important laboratory for AI governance because so many of the frontier developers are based here,” Ranjit said.
Ranjit said she and her colleagues are “cautiously optimistic” about the possibilities AI regulation presents and she hopes that an “ecosystem of evaluators who come from diverse backgrounds” will be able to effectively regulate AI.
“There is no single company, government agency, or evaluator that is going to have all of the necessary expertise to understand general purpose AI systems,” Ranjit said.
Matthew Bevins, a sophomore majoring in AI for business, said he believes that while tech companies need to slow AI development, they cannot halt advancement completely.
“I think a more advanced AI would make the job market a lot more difficult, but the thing is, AI is already extremely advanced, and there are a lot of jobs that it could already replace,” Bevins said. “The real issue when it comes to unemployment is adoption, and most companies have been a lot slower to adopt than previously expected.”
Bevins also said he believes there is no point in focusing solely on the downside.
“I feel like it’s overblown, and people shouldn’t have the mindset going through life that they’re screwed because AI is going to take over the world,” Bevins said.
Bevins said he believes Trump is viewing the issue solely through an international relations lens, since he wants to stay ahead of China in the AI race, but the bigger picture matters.
“It’s about protecting the interests of everyday Americans, protecting them from having their data used in a way that they don’t want and protecting them from high levels of unemployment that come from AI,” Bevins said.
Edited by JohVonne Roberts, Ally Nakamura and Andriana Yatsyshyn.