California lawmakers have passed a controversial bill that could significantly impact the development and use of artificial intelligence (AI). The legislation, known as SB 1047, seeks to ensure the safety of advanced AI models in everyday life.
Tech giants like Google, Microsoft, and OpenAI have raised concerns about the bill, arguing that it could stifle innovation and drive AI companies out of the state.
However, proponents, including Tesla CEO Elon Musk, believe that the regulations are necessary to protect the public from potential risks associated with powerful AI.
So, what does this mean for Californian residents and the world beyond? Let’s take a closer look.
California’s track record in safety bills
The new AI bill is the first to pass a parliamentary motion in the United States, and now just awaits the signature of state governor Gavin Newsome before going into law.
It’s major news for AI safety advocates, but maybe not much of a surprise for those who know about the state’s history with online safety.
For a start, it’s one of the country’s fiercest critics of internet betting: playing online casino games is outlawed there, and sports betting is extremely restricted. The increasing number of problem gamblers across the US suggests this won’t change any time soon.
Since the dawn of AI, California has also led American efforts to regulate it. The California Privacy Protection Agency (CPPA) proposed plans to regulate automated tools and facial recognition in November 2023. These suggested that companies should advise their customers that they use AI tech in the creation of their product or service and give California residents the right to stop the use of their personal data by AI-powered models. Employers would also be required to notify job applicants if a decision regarding their employment was based on AI.
The new SB1047 bill, though, is a major next step in this process. The biggest piece of legislation yet, it could pave the way for wider-reaching laws in the future.
What’s in the new bill?
Central to the new bill is the introduction of safety testing for AI models that meet certain criteria, including those costing over $100 million to develop or that need a high level of computing power.
This will mean that powerful models are tested closely for risks before being released to the public. The bill will also require developers to install a “kill switch” within the model that shuts it down should it start to malfunction or pose a threat.
Independent auditors will also assess the models even after they pass the safety process and appear on the market. Whistleblowers who report unethical or illegal AI practices will also receive protection should they report any harmful activity or malfunction from within the tech.
However, the bill has also met significant resistance from both sides of the AI debate.
The big debate
The debate over SB 1047 is centered on just how risky AI is. The bill’s supporters argue that it is necessary to protect the public from potential AI-related risks, with some even suggesting that the legislation doesn’t go far enough in this respect.
Opponents say it could stifle innovation and drive AI companies out of California. Tech giants like Google, Microsoft, and OpenAI believe the bill will have a negative impact on their research and development efforts.
Others, like Time Magazine, say that the bill makes the mistake of regulating the technology rather than its applications.
Yet the likelihood is that the bill is the state’s best chance at treading the fine line between the two viewpoints. It aims to promote AI innovation with a “safety harness” on, something that many AI critics have championed in recent years.
How SB 1047 could affect the future of AI regulation
If the bill becomes law, it could set a precedent for other jurisdictions to follow, most notably in other US states and in major tech centers like London, Tokyo, and Brussels.
World leaders have frequently mentioned the need to control the planet’s most disruptive tech since the advent of widespread internet access three decades ago. Some scientists even estimate the chances of AI destroying the world to be at around 5%, a calculation that has driven authorities to take action.
That said, the Californian bill is just one small step toward adequate regulation. For regulators the battle has only just started, with mountains of paperwork and legal red tape blocking the road ahead of them.
Also Read: Optimizing Customer Service With AI-Powered Language Tools
