
Fri Sep 06 15:57:00 UTC 2024: – California’s AI bill, Senate Bill 1047, aims to regulate advanced AI systems by requiring testing for models costing a minimum of $100 million to train for potential harm and implementing guardrails for mitigation.
– The bill has garnered support from AI experts like Geoffrey Hinton and Yoshua Bengio, but has also faced criticism from Rep. Nancy Pelosi and OpenAI.
– Analysts like Avivah Litan see both sides of the issue, acknowledging the need for regulation but also expressing concerns about potential negative impacts on small businesses and innovation.
– Some argue that focusing on regulating applications rather than technology itself, as seen in the EU AI Act, may be a more effective approach.
– The bill also includes whistleblower protections, but concerns remain about compliance costs for small businesses and outdated thresholds for AI models.
– Other states like Colorado and Connecticut are also advancing AI bills, creating a complex regulatory landscape that some argue should be handled at the federal level.
– While a federal AI standard is unlikely to be voted on soon, multiple AI bills have been proposed at the federal level without success.
California AI bill sets guardrails that draw criticism
California AI bill sets guardrails that draw criticism