The California Approach
SB 1047-style AI regulation poses a threat to US hegemony in the AI race. Some particular concerns underscored by TPPF research include: Impossibly high pre-deployment standards, such as the assurance that there is zero possibility for covered models to have hazardous capabilities. Enmeshing AI into state functions with an explicit goal of “fostering equitable innovation.”
Definitional ambiguity. The goal of good public policy is to create certainty and compliance, whereas SB 1047 hastily defines terms in a manner that broadens scope and application outside of professed goals. Process woes. California lawmakers ostracized key stakeholders from providing input in the legislative process, myopically focused on eliminating potential harms at all costs without regard to trade-offs and technical feasibility. TPPF recognizes this as the “shiny object syndrome” that is driving many left-leaning states. The reflex is to regulate without regard for national security, economic, and entrepreneurial ramifications. Driving innovative companies out of CA is a feature, not a bug, and states like Texas welcome them with open arms and regulatory certainty.