Menu

Search

  |   Law

Menu

  |   Law

Search

OpenAI May Avoid California Incorporation Amid Stricter AI Data Disclosure Law

OpenAI faces challenges due to California’s new AI data disclosure law, complicating expansion plans. Credit: EconoTimes

California's new AI Training Data Transparency Bill, signed into law by Governor Gavin Newsom, could hinder OpenAI's commercial plans. The legislation mandates detailed disclosures of datasets used in AI model development, potentially deterring OpenAI from incorporating its for-profit arm in the state.

California’s New AI Transparency Law Poses Challenges for OpenAI's Commercial Expansion Strategy

The State of California has recently taken a significant stride toward ensuring that the development and operation of generative AI models are conducted with greater transparency. Nevertheless, this action poses competitive risks for OpenAI, a pioneer in the field, as it pursues an increasingly commercial strategy. Consequently, OpenAI would elect to establish its for-profit subsidiary in California in light of the state's recent regulations.

In recent months, OpenAI has transitioned to a more commercial trajectory under the stewardship of CEO Sam Altman. Altman's focus on profitability has been a source of disagreement within the organization, reportedly leading to internal discord toward the conclusion of 2023. According to Wccftech, this change was further underscored by the disclosure that Altman was negotiating to acquire a 7 percent ownership stake in OpenAI as part of a new $6.5 billion funding round. This transaction would value the company at $150 billion and Altman's stake at approximately $10.5 billion.

Altman has suggested that OpenAI be transformed from a non-profit to a public benefit corporation in conjunction with this transition to ostensibly allocate profits to public welfare. Despite this commercial pivot, OpenAI's non-profit arm would persist and retain a minority stake in the for-profit arm.

The potential challenge to this commercial transition is the new law in California, which Governor Gavin Newsom signed on September 28. The AI Training Data Transparency Bill (AB 2013) requires all AI models utilized within the state to provide a comprehensive summary of the datasets used in their development. The required information encompasses the disclosure of dataset sources, owners, utilization, and whether the data is public domain or protected by intellectual property rights. In addition, developers must disclose the following: whether datasets were purchased, licensed, or contain personal data, the method by which the datasets were processed, and the use of synthetic data.

OpenAI Faces Hurdles as California's AI Transparency Law Increases Compliance Costs and Risks

OpenAI, which has been notoriously secretive regarding its proprietary datasets and training methodologies, may encounter substantial challenges due to this degree of transparency. For instance, the technical report for GPT-4 did not provide significant information regarding the model's training or data sources. OpenAI has recently warned users against attempting to reverse-engineer its o1 model and has even threatened to ban those who tried to trace its decision-making process.

The company may be discouraged from incorporating its for-profit subsidiary in California due to the increased cost of compliance with California's transparency requirements due to OpenAI's evolving commercial interests. While OpenAI has advocated for other California regulations, including AB 3211, which mandates the labeling of AI-generated content, it has been critical of SB 1047, which mandates that developers undertake safety testing on AI models.

To provide context, OpenAI has recently released two significant models: the o1 model, which emulates human problem-solving through a train-of-thought approach, and GPT-4o, a versatile model capable of processing text, audio, and visual inputs. Despite its superior accuracy, OpenAI's business strategy is further complicated by the fact that the o1 model is substantially slower and more expensive than GPT-4o.

The company may encounter additional obstacles as it attempts to reconcile innovation with compliance in the AI sector in response to California's recent regulations.

  • Market Data
Close

Welcome to EconoTimes

Sign up for daily updates for the most important
stories unfolding in the global economy.