OpenAI Presents Its Favorite Proposal for AI Regulation in a New 'Framework'

OpenAI has presented an "economic plan" for artificial intelligence, a constantly evolving document that outlines the policies the company wishes to develop in collaboration with the U.S. government and its allies. This plan, which includes an introduction by Chris Lehane, OpenAI's Vice President of Global Affairs, emphasizes the need for the United States to attract billions in investment for essential resources like chips, data, energy, and talent, in order to "succeed in artificial intelligence." Lehane states that while some countries marginalize the economic potential of AI, the U.S. government has the opportunity to facilitate the country's global leadership in innovation while also protecting national security.

OpenAI has urged the U.S. government to take more concrete actions to regulate AI, arguing that current regulation is fragmented, as the federal government has largely delegated these functions to the states. In 2024, nearly 700 bills related to AI were introduced in various state legislatures, some of which conflict with each other. For example, Texas's Responsible AI Governance Act imposes strict liability requirements on developers of open-source AI models.

OpenAI's CEO, Sam Altman, has also criticized existing federal laws, such as the CHIPS Act, which aimed to revitalize the semiconductor industry in the U.S. by attracting local investments. In a recent interview, Altman remarked that the CHIPS Act "has not been as effective as we had hoped," and he believes there is a "real opportunity" for the Trump administration to carry out more effective initiatives. Altman pointed out that in the U.S., building infrastructure has become excessively complicated, which is problematic for leadership in AI.

To develop and maintain the data centers needed for AI, OpenAI's plan suggests a "drastic" increase in federal spending on energy and data transmission, as well as a significant expansion of new energy sources, including solar projects, wind farms, and nuclear energy. The company, along with its competitors, has previously supported nuclear energy projects, arguing that they are essential to meet the electrical needs of future server centers.

In the short term, the plan recommends that the government develop "best practices" for model implementation, streamline the AI industry's relationship with national security agencies, and establish export controls that allow sharing models with allies while limiting exports to adversarial countries. It is also proposed that the government share national security-related information that may affect the AI industry and help providers assess the risks of their models.

OpenAI already has some government departments as partners, and if the plan is adopted, it could establish further collaborations. It currently maintains agreements with the Pentagon for cybersecurity projects and has collaborated with the defense company Anduril to provide AI technology to combat drone attacks. According to the plan, it is suggested that recognized standards be developed by other nations on behalf of the U.S. private sector, although no mandatory rules are proposed.

The draft also addresses copyright issues related to AI, arguing that developers should have the ability to use "publicly available information," including copyrighted material, to create models. OpenAI contends that training AI models without using protected materials would be "impossible" and has faced lawsuits from creators who claim the company used their works without authorization. In this regard, the plan states that if the U.S. and like-minded nations do not address this imbalance through appropriate measures, other countries will exploit this content for their own economic benefit.

It is unclear which aspects of OpenAI's plan will effectively influence future legislation. However, the document indicates that the company is committed to being a relevant player in the formulation of a unified AI policy in the U.S. During the first half of last year, OpenAI significantly increased its lobbying expenses and has brought in former government leaders to its team, in addition to supporting legislative initiatives in favor of a federal regulatory body for AI and grants for research and development.