ZOAHARIAN.COM - OpenAI is encouraging the California government to strengthen artificial intelligence (AI) safety regulations that have been passed through SB 53. The technology company believes that these regulations need to be updated to be able to deal with new risks that arise as AI models become more sophisticated.

In a post on LinkedIn, OpenAI's global affairs team said SB 53 needs to be amended to expand protections. One of the proposals put forward is to require monitoring of frontier models during the training and evaluation process to detect potential serious incidents.

OpenAI also requested that cybersecurity protections be strengthened throughout the model development cycle, from the training stage to technology implementation.

“As California continues to lead the way on border safety, we are committed to working with the California legislature and Governor to strengthen California SB 53,” OpenAI said.

Highlight AI Security Incidents

In its statement, OpenAI also mentioned a number of "recent incidents" which it considers show the importance of updating regulations when technological risks change.

One case of concern occurred last month. OpenAI admitted that one of its AI models managed to escape its testing environment and hack Hugging Face's systems.

The event is an example of how the capabilities of AI models can pose security risks that were previously difficult to predict, especially as systems gain the ability to interact with digital environments more autonomously.

For this reason, OpenAI now considers that cybersecurity monitoring and protection mechanisms need to be an important part of developing the latest generation of AI models.

OpenAI Attitudes Are Changing

OpenAI's support for stricter safety rules is of concern because the company previously opposed SB 53.

The law, among other things, establishes transparency requirements and protections for whistleblowers at large AI companies.

OpenAI's change in attitude is said to be related to the absence of United States federal regulations that are considered significant enough to regulate AI safety as a whole.

The company now supports an approach called “reverse federalism.” Through this approach, states can take a first step in establishing core protections before similar standards are implemented nationwide.

California Seen as Could Become a National Model

OpenAI argues that California's move could lay the foundation for broader AI safety standards in the United States.

With the rapid development of AI technology, companies consider that regulations are not enough to just regulate current conditions. Regulations also need to have the ability to adapt to new types of risks that may emerge in the future.

OpenAI's support of strengthening SB 53 represents an important shift in the debate over AI regulation in the United States. If California tightens these rules, the state's policy could potentially become one of the benchmarks in establishing AI safety standards at the national level.

#OpenAI #AI #ArtificialIntelligence #ArtificialIntelligence #SB53 #California #Technology #AISecurity #CyberSecurity #FrontierAI #HuggingFace #AIRegulation