Fordel StudiosCalifornia passes controversial bill regulating AI model training
What Happened
As the world debates what is right and what is wrong about generative AI, the California State Assembly and Senate have just passed the Safe and Secure Innovation for Frontier Artificial Intelligence Models Act bill (SB 1047), which is one of the first significant regulations for AIs in the United S
Fordel's Take
California's SB 1047 passed both chambers. It mandates safety testing, killswitches, and incident reporting for any AI model trained above roughly $100M in compute cost — before that model ships.
This targets OpenAI and Anthropic directly, not developers calling their APIs. Your RAG pipeline or fine-tuned Llama 3 deployment doesn't qualify as a "frontier model" under this bill. Assuming your API wrapper inherits compliance obligations is how teams waste legal budget on problems they don't own.
Teams deploying inference via Anthropic or OpenAI APIs carry zero burden here. Fine-tuners on open models below the compute threshold are also clear.
What To Do
Do a one-page legal scope check against the $100M compute threshold instead of blanket compliance prep because most production AI systems are RAG or fine-tune stacks that don't qualify.
Cited By
React
Get the weekly AI digest
The stories that matter, with a builder's perspective. Every Thursday.