Criticism Over Focus on Large Models with Investment Exceeding 100 Million Dollars
The state of California, the hub of advanced technology development in the United States, has put the brakes on artificial intelligence (AI) regulation legislation.
On the 29th (local time), Governor Gavin Newsom of California exercised his veto power against the regulatory bill 'SB 1047,' which strengthens the responsibilities of AI development companies. This came just over a month after the state legislature passed the bill, with the veto issued one day before the signing deadline.
Earlier, Democratic State Senator Scott Wiener proposed the bill, stating, "There is a need to create laws to protect the public before AI becomes uncontrollable." The bill mandates safety testing for large language models (LLMs) with investments of at least $100 million (approximately 130 billion KRW) and introduces a 'kill switch' to forcibly shut down AI that operates beyond control. It also includes provisions allowing the state attorney general to file lawsuits if companies fail to comply with the regulations.
However, Governor Newsom explained the reason for his veto, saying, "The regulation focuses only on the largest and most expensive AI models," and "It attempted to regulate based solely on the size and cost of AI models without considering whether those models are actually used in dangerous situations." He pointed out that smaller AI models can also pose risks when handling sensitive data such as power grids or medical records, while large models are sometimes used for relatively low-risk tasks like customer service.
Governor Newsom announced plans to require the state government to investigate the potential threats that generative AI could pose to major infrastructure within California, and to conduct risk and capability analyses of major AI models led by Stanford University professor Fei-Fei Li, known as the 'godmother of AI.'
With Governor Newsom's veto, big tech companies can breathe a sigh of relief. Google, Meta Platforms, Microsoft (MS), OpenAI, and others have opposed the bill, arguing that it imposes vague standards under the guise of safety. Many small and medium-sized enterprises have also expressed concerns that the bill could discourage big tech from opening AI models to the public, thereby hindering innovation and destroying the startup ecosystem.
Meanwhile, on the 28th, Governor Newsom signed and enacted an amendment to the privacy law to prevent the misuse of people's brain information by neurotechnology companies. This comes amid accelerated development of devices for meditation, concentration enhancement, and depression treatment, raising concerns about the potential misuse of individuals' brain data. Under this law, individuals will be able to request deletion or modification of their neural data collected by companies.
© The Asia Business Daily(www.asiae.co.kr). All rights reserved.
