Since AI began to gain historical attention in 2022, the debate on how to regulate the industry has been growing at a similar pace. Many startup founders believe that regulating the industry now could stifle innovation. However, others believe that waiting to implement policies and regulations could have a negative impact on the industry in the future.
Helen Tonner, director of strategic and basic research grants at the Center for Security and Emerging Technologies and a former OpenAI executive, understands the AI industry's fears about regulation. But at TechCrunch's StrictlyVC event in June, Tonner told an audience that AI companies shouldn't fear regulation now, but rather that it could come suddenly later.
She added that if the U.S. government delays implementing regulations now, the AI industry could suffer the same fate as sectors like cryptocurrency and social media: If an incident occurs, Congress will likely quickly pass legislation to regulate the industry, but it could be more heavy-handed or less thoughtful than if it had been enacted before something happened.
“Some of the smarter and more thoughtful actors that I've seen in this space are trying to say, 'Okay, what are some fairly light-touch, common-sense guardrails that we can put in place now to reduce the occurrence of future crises, to reduce their severity, and basically reduce the likelihood that we'll need some kind of quick, ill-considered response later on?'” she said.
Toner said that regardless of how startups and venture capitalists feel about the possibility of AI regulation, they should be proactive in talking about it. Companies of all sizes and stages should be talking to people and lobbyists in Washington about how they operate, what different regulations might look like, and how it would affect their business. That way, they can have a say in shaping the regulations that work for them.
“I think there's really value in thinking of it as a productive collaboration of different people with different expertise, bringing your own expertise to the table,” Toner said. “If you think of it as something to be ignored, or as an adversarial battle where you're trying to stop the other side from doing as much as possible, I think it's less productive.”
Smaller startups and entrepreneurs looking to expand into the AI space can also start preparing for potential regulation, Tonner said.
“[Companies should be] “They need to track what they're training, how they're training it, how they're testing it, what they're doing with the data, how they're updating it, which models they're deploying at what point in time in what products,” she says. “What I've heard from people inside these large companies, really well-respected companies, is that they often don't even have proper records of things like what model mixes are currently running certain products.”
But Congress has been holding hearings on the possibility of regulating AI, and while Toner is confident it will happen at some point, he doesn't think it will happen anytime soon, regardless of how this year's election cycle goes.
“If I had to bet on this, I would bet that there won't be any major new changes at the federal level,” Toner said, “and it's a question of what else happens around there.”