Key Points
- California revises AI legislation after pushback from top AI companies.
- Senator Scott Wiener addresses concerns, including those from Anthropic.
- AI pioneers Geoffrey Hinton and Yoshua Bengio back the revised bill.
- The bill targets frontier AI models with significant computing power.
California’s efforts to regulate artificial intelligence (AI) have hit a critical juncture, with lawmakers revising key provisions of a controversial bill following significant backlash from industry leaders.
The Safe and Secure Innovation for Frontier Artificial Intelligence Models Act (SB 1047), spearheaded by State Senator Scott Wiener (D), has been at the center of a heated debate, with some of the most prominent names in the AI industry voicing concerns about its potential impact on innovation.
Industry Concerns Lead to Legislative Revisions
The bill, originally designed to impose strict regulations on advanced AI models, has faced resistance from leading AI companies, including Anthropic, a company founded by former OpenAI staff.
Congressional Representative Zoe Lofgren (D-CA-18) also expressed concerns, warning that the legislation might drive AI companies out of California, a state that has become synonymous with AI innovation.
In an August 7 letter to Senator Wiener, Lofgren articulated her fears that the bill could stifle California’s thriving AI industry without offering clear public benefits. “There is a real risk that companies will decide to incorporate in other jurisdictions or simply not release models in California,” she cautioned.
California’s significance in the AI industry cannot be overstated. According to Governor Gavin Newsom’s executive order from last September, the state is home to 35 of the top 50 AI companies globally.
This makes the legislation’s implications far-reaching, not just for California but for the AI industry at large.
In response to the growing concerns, Wiener announced on Thursday that the bill had been revised to address the core issues raised by companies like Anthropic.
“While the amendments do not reflect 100 percent of the changes requested by Anthropic…we accepted several very reasonable amendments proposed,” Wiener said. He emphasized that the revised bill addresses the industry’s primary concerns while maintaining its regulatory intent.
California trims AI safety bill amid fears of tech exodus https://t.co/wFX1iqd4Jg pic.twitter.com/Cvry9q0952
— Tech News Tube (@TechNewsTube) August 17, 2024
AI Pioneers Voice Support for the Bill
Despite the pushback, the bill has garnered support from several AI pioneers, including Geoffrey Hinton, emeritus professor of computer science at the University of Toronto, and Yoshua Bengio, professor of computer science at the University of Montreal.
Both Hinton and Bengio are considered luminaries in the field of AI, having contributed foundational research that underpins much of today’s AI technology.
Hinton, who has been a vocal advocate for responsible AI development, praised the revised legislation for its balanced approach.
“Forty years ago, when I was training the first version of the AI algorithms behind tools like ChatGPT, no one – including myself – would have predicted how far AI would progress,” Hinton said in a statement. “Powerful AI systems bring incredible promise, but the risks are also very real and should be taken extremely seriously.”
Bengio echoed Hinton’s sentiments, underscoring the importance of legislation that both fosters innovation and mitigates risks. “California is a natural place for that to start, as it is the place this technology has taken off,” he noted.
The bill specifically targets “frontier models,” a term referring to state-of-the-art AI systems that require significant computational resources to develop.
These models, which involve more than 10²⁶ integer or floating-point operations and cost over $100 million to train, represent the cutting edge of AI technology.
The Future of AI Regulation
The cost of training these frontier models is escalating rapidly. According to Anthropic CEO Dario Amodei, AI models currently cost around $100 million to train, with some models in development that could cost up to $1 billion.
He predicts that within the next few years, these figures could balloon to $10 billion or even $100 billion as chip and algorithm improvements continue.
Amodei’s comments highlight the immense resources required to develop the next generation of AI, as well as the potential power of these models.
“There is in my mind a good chance that by that time we’ll be able to get models that are better than most humans at most things,” Amodei said, reflecting on the rapidly approaching capabilities of AI systems.
As the revised SB 1047 moves closer to a vote in the California State Assembly, the debate over AI regulation continues to underscore the delicate balance between fostering innovation and ensuring public safety.
California’s role as a leader in the AI industry makes the outcome of this legislation particularly significant, setting a precedent that could influence AI policy worldwide.
You May Also Like This Post
