OpenAI has laid out a new “economic blueprint” for the artificial intelligence industry that aims to ensure America remains at the forefront of innovation.

In a blog post today that introduced the new document, the company stresses that it believes in America “because America believes in innovation,” and tries to position itself as a champion of tech leadership in the country, while also raising the alarm over the prospect of Chinese rivals catching up.

The company appears to be trying to create a sense of urgency regarding the need for more funding. In the document, it stresses that there is about $175 billion in global investment capital that has been earmarked for AI startups. And it warns that if American companies don’t attract most of that money, it could instead flow into Chinese-backed companies, expanding that country’s leadership in the AI industry at the expense of the U.S.

OpenAI is calling for stricter export controls on frontier AI models to prevent such a scenario, saying that access should be limited to the U.S. and allied nations only, similar to what the U.S. government is doing with computer chip technology. That follows a decision last year by the company to start restricting access to its own technology.

AI economic zones

At the same time, OpenAI is pushing for a more unified approach to regulate the AI industry in America, saying this should be done at the national level, rather than relying on a patchwork of state rules. Its strategy for regulation is based on four key pillars it says are critical to AI leadership – chips, data, energy and talent.

In addition to regulation, the company also outlined some ambitious plans for building the country’s AI infrastructure. In its document, it calls for the creation of “AI Economic Zones,” which sound similar to the U.K.’s proposed “AI growth zones” that were announced by its Prime Minister Keir Starmer yesterday.

Within these economic zones, it calls for more lenient rules that would enable the construction of new solar, wind and nuclear power plants to be fast-tracked. It also wants to link these zones through a “National AI Infrastructure Highway” that would connect regional power and communications networks.

Each of the regions would specialize in a different aspect of AI, based on its local economy, OpenAI said. For instance, it suggests that Kansas might become a hub for AI-powered agriculture, while Pennsylvania or Texas could become proving grounds for AI energy innovation. It envisions that these states could work more closely with local industries to ensure wider adoption of AI technologies, accelerating innovation.

Analyst Rob Enderle of the Enderle Group told SiliconANGLE that the idea of AI economic zones is interesting, but could benefit from being tweaked, so that zones are created where there is already a critical mass of developers and common interest in focus areas.

“Existing universities that have the relevant focus already might be a better core for these efforts, instead of trying to do it from scratch,” Enderle said. “But this also seems like an old way of thinking, for today collaboration occurs across broad geographic areas and across borders, so outside of universities, I’m not sure the idea of having a concentrated focus area makes much sense.”

In addition, the company says it might be necessary for AI companies to have access to intelligence information, and to do that they should build more secure data centers with classified computing clusters, where they can keep their frontier models safe.

OpenAI said the blueprint is the first step in a wider “Innovating for America” initiative that aims to ensure AI technology will ultimately benefit all Americans, regardless of their background. On Jan. 30, Chief Executive Sam Altman (pictured) will travel to Washington D.C. to meet Trump administration officials and detail its plans for AI-driven economic growth.

Crackdown on foreign AI projects

OpenAI also discussed its stance on intellectual property in the document, and it’s here that its ideas begin to get more controversial. It argues that AI systems should be able to learn from “universal, publicly available information” without any restrictions. That appears to be an effort to bolster its argument that scraping the internet falls under “fair use” doctrine.

However, while defending its own rights to scrape the web for more data, OpenAI also calls for a crackdown on “unauthorized digital replicas,” including some foreign AI firms that it says “make no effort to respect or engage with the owners of IP rights.”

Enderle said the challenge around regulating this kind of practice is that humans themselves do exactly the same thing, albeit at smaller scales. “This makes it difficult to create restrictions that wouldn’t apply to humans too, and we are nowhere near sorting this out,” he said.

OpenAI believes there’s a need to do something though. It argues that if the U.S. and its allies don’t attempt to address this issue by initiating sensible measures designed to enhance AI in the long-term, “the same content will still be used for AI training elsewhere, but for the benefit of other economies.”

In other words, OpenAI appears to be arguing that it’s OK for it and other American AI companies to help themselves to any and all public data, but the government should make efforts to prevent those from less friendly nations from doing the same.

Enderle said it may make more sense to try and cooperate with its rivals. He argued that while there are justified concerns around China due to its reputation for stealing intellectual property, the U.S. could find itself falling behind if it shuts the door to any cross-border collaboration.

“Breakthroughs can occur in any country with a critical mass of developers and mature AI development engines, and such a stance could prevent U.S. companies from accessing those developments,” he said.

Photo: TechCrunch/Flickr