SB 1047: ‘safe & secure’ really means ‘see you in court’

Capitol Weekly version posted here

By Kyla Christofferson Powell

OPINION – In May of this year, Governor Newsom warned, “If we over-regulate, if we overindulge, if we chase the shiny object, we could put ourselves in a perilous position.”

Sadly, Senator Scott Wiener’s SB 1047 bill falls into the trap Governor Newsom warned about. This bill would put California’s economy in a difficult position by regulating an artificial intelligence (AI) industry that policymakers have yet to understand. SB 1047 is chasing the shiny object of a new sector that grabs headlines instead of thoughtfully considering the long-term adverse effects of creating a solution in search of a problem.

SB 1047 is intended to set up guardrails for AI, but instead it offers up platitudes like “safe and secure” that don’t have any clear meaning. Even worse, it creates new punishments that would make AI the next frontier in litigation instead of the advent of a new technology. SB 1047 would jeopardize California’s leadership role in AI by targeting AI developers.

Instead of fostering the development of this new and growing industry that could have broad and positive impacts on our economy, SB 1047 empowers California’s Attorney General to take civil action against AI developers and seek substantial penalties if they make a mistake in the AI development process.

The “gotcha process” of conflicting and confusing requirements in SB 1047 creates significant regulatory uncertainty, imposes high compliance costs, and needlessly creates substantial liability risks for failing to predict a harmful use of their models by others.

The bill even offers jury instructions to assist lawyers in preparing legal cases against innovators and AI developers. At its most severe, it authorizes the shutdown and destruction of AI models –even those that could potentially save lives in areas such as drug discovery, medical diagnostics, and disaster response.

A better approach for California, a global leader in technology for decades, is to engage our world-class resources.  First, we shouldn’t legislate until we fully understand AI technology’s possible uses and risks. Instead, we should allow time for state agencies to implement Governor Newsom’s Executive Order on AI and complete their assessment of AI technology. That process, involving high-level state technology personnel, should be the basis for developing policies.

Second, we should also involve technology experts from California’s colleges and universities in developing guardrails because they have different areas of expertise and collectively form the world’s best minds on technology and policy. Third, AI developers, the technology sector developing AI models, and those currently using AI need to be involved in developing laws and regulations that will affect their jobs and businesses.

There is a lot at stake in regulating artificial intelligence. This technology is too important and promising for the future of California’s economy to put at risk misguided legislation created in a vacuum of understanding. The legislature must reject SB 1047 and, in its place, make the path that achieves the balance between leadership on AI and fair and effective guardrails.

Powell is the President & CEO of the Civil Justice Association of California, a source of expertise in legal reform and advocacy for almost half a century.