California's AI safety law proves regulation and innovation can coexist

Featured Image

California’s SB 53: A New Era of AI Regulation

California Governor Gavin Newsom recently signed the AI Safety and Transparency Bill, known as SB 53, into law. This legislation is seen as a significant step in balancing innovation with safety in the rapidly evolving AI landscape. Adam Billen, vice president of public policy at Encode AI, emphasized that state regulation does not have to stifle AI progress. Instead, it can serve as a framework for ensuring that AI technologies are developed responsibly.

At its core, SB 53 is a first-in-the-nation bill that requires large AI labs to be transparent about their safety and security protocols. Specifically, it focuses on how these labs prevent their models from posing catastrophic risks, such as being used for cyber attacks on critical infrastructure or the development of bio-weapons. The law also mandates that companies adhere to these protocols, which will be enforced by the Office of Emergency Services.

Billen pointed out that many companies already engage in safety testing and release model cards. However, he noted that some companies may begin to cut corners under competitive pressure. This is where legislation like SB 53 becomes essential. For instance, OpenAI has publicly stated that it may "adjust" its safety requirements if a rival AI lab releases a high-risk system without similar safeguards. Billen argues that policy can enforce existing safety promises, preventing companies from compromising on safety standards.

While public opposition to SB 53 was relatively muted compared to its predecessor, SB 1047, which Newsom vetoed last year, the rhetoric in Silicon Valley and among most AI labs has been that any AI regulation is anathema to progress. This sentiment has led to efforts by companies like Meta, venture capital firms like Andreessen Horowitz, and individuals like OpenAI president Greg Brockman to invest heavily in super PACs supporting pro-AI politicians in state elections. Additionally, they pushed for an AI moratorium that would have banned states from regulating AI for 10 years.

Encode AI, along with over 200 organizations, worked to strike down this proposal. However, the fight is not over. Senator Ted Cruz, who championed the moratorium, is attempting a new strategy to achieve the same goal through federal preemption of state laws. In September, Cruz introduced the SANDBOX Act, which would allow AI companies to apply for waivers to temporarily bypass certain federal regulations for up to 10 years. Billen also anticipates a forthcoming bill establishing a federal AI standard, which could override state laws.

Billen warned that narrowly scoped federal AI legislation could undermine federalism, particularly for the most important technology of our time. He emphasized that while the AI race with China matters, killing state bills focused on issues like deepfakes, transparency, algorithmic discrimination, children’s safety, and governmental use of AI is not the way to go. He argued that such bills do not hinder the U.S. from beating China in the AI race.

Legislative proposals like the Chip Security Act aim to prevent the diversion of advanced AI chips to China through export controls and tracking devices. The existing CHIPS and Science Act seeks to boost domestic chip production. However, some major tech companies, including OpenAI and Nvidia, have expressed reluctance or opposition to certain aspects of these efforts, citing concerns about effectiveness, competitiveness, and security vulnerabilities.

Nvidia, for example, has a strong financial incentive to continue selling chips to China, which historically represented a significant portion of its global revenue. Billen speculated that OpenAI might hold back on chip export advocacy to maintain good relations with crucial suppliers like Nvidia. There has also been inconsistent messaging from the Trump administration, which initially expanded an export ban on advanced AI chips to China but later reversed course, allowing Nvidia and AMD to sell some chips to China in exchange for 15% of the revenue.

Billen noted that while there is movement towards bills like the Chip Security Act that would put export controls on China, there continues to be support for narratives that aim to kill state bills focusing on light but necessary regulations. He added that SB 53 is an example of democracy in action, showcasing how industry and policymakers can work together to create a bill that everyone can agree on. While the process may be messy, it reflects the foundation of the country's economic system, and he hopes it continues successfully.

“I think SB 53 is one of the best proof points that that can still work,” he said.

Post a Comment for "California's AI safety law proves regulation and innovation can coexist"