California State Assembly Passes Landmark AI Safety Bill
The Push for AI Safety Regulations
The California State Assembly has passed the Safe and Secure Innovation for Frontier AI Models Act (SB 1047), a landmark step towards regulating artificial intelligence in the United States. The bill mandates that AI companies implement crucial measures prior to training advanced models, such as enabling swift shutdowns and protecting against unsafe post-training modifications.
Essential Provisions of SB 1047
- Mandatory testing for catastrophic safety risks
- Immediate shutdown capabilities for models
- Protection against unsafe modifications post-training
According to Senator Scott Wiener, the legislation seeks to ensure that AI labs conduct safety tests to mitigate foreseeable risks. Despite receiving criticism from companies like OpenAI and Anthropic, the bill has been modified to address concerns surrounding small developers and enforcement.
The Next Steps for AI Safety in California
Following its passage in the Assembly, the AI safety bill will move on to the State Senate for a vote that is anticipated to succeed. Once cleared, it will be presented to Governor Gavin Newsom, who will decide on the bill's future by the end of September, as stated by The New York Times.
This article was prepared using information from open sources in accordance with the principles of Ethical Policy. The editorial team is not responsible for absolute accuracy, as it relies on data from the sources referenced.