OpenAI Calls on U.S. Government to Feed Its Data Into AI Systems

OpenAI wants you to think of AI like a car. Europe invented the car, but heavy regulations prevented its widespread adoption there. In laissez-faire America, the car dominated the culture. OpenAI wants the U.S. to do that again. On Monday the company behind ChatGPT published AI in America: OpenAI’s Economic Blueprint, a whitepaper that calls on Washington to let AI determine the country’s future.

AI in America is a slight 15-page document with an AI-generated photo on its cover that shows an architect’s desk overlooking a futuristic cityscape. The picture and 15 pages seem fine at first glance. But like so much of the stuff attached to AI, both the image and the outline for economic prosperity seem vague and grotesque the more you inspect them. The picture’s coffee cup has no handle. The words written on the pages of the picture look like unreadable smudges. The economic blueprint contains calls to action that ask the government to turn over public secrets to large private companies.

The more you look, the more things fall apart. OpenAI’s Economic Blueprint is a call for a lightly regulated AI future where government-collected data, both state secrets and public information, is fed into its vast and hungry machines.

The first thing OpenAI wants you to know is that AI is very important and very scary. “AI is too powerful to be led and shaped by autocrats, but that is the growing risk we face, while the economic opportunity AI presents is too compelling to forfeit,” reads an opening note from OpenAI vice president of global affairs Chris Lehane. “Shared prosperity is as near and measurable as the new jobs and growth⁠(opens in a new window) that will come from building more AI infrastructure like data centers, chip manufacturing facilities, and power plants.”

And how should America accomplish such ambitious goals? By sharing as many of its secrets as possible with AI companies. “As appropriate, share national security-related information and resources that it alone maintains—such as briefings on security threats to the industry and high-level results of testing US and non-US AI models—with US AI companies pursuing advanced research,” the economic blueprints says.

It goes further. OpenAI always wants the feds to share their “unique expertise with AI companies, including information about how to secure their IP against industrial security threats and mitigate potential cyber, chemical, biological, radiological and nuclear CBRN and other risks that are heightened by increasingly powerful models.”

And of course, there’s all that wonderful data just sitting around waiting to be scanned. “A lot of government data is in the public domain. Making it more accessible or machine-readable could help US AI developers of all sizes, especially those working in fields where vital data is disproportionately government-held,” the blueprint says. “In exchange, developers using this data could work with government to unlock new insights that help it develop better public policies.”

The economic blueprint also notes that “infrastructure is destiny” and that the U.S. is in a unique position to create jobs and get ahead of China. All it has to do is focus on building infrastructure for AI systems, if not people. “In the AI era, chips, data, energy and talent are the resources that will underpin continued US leadership, and as with the mass production of the automobile, marshalling these resources will create widespread economic opportunity and reinforce our global competitiveness,” it says.

What does that mean? “Seizing the moment and building the infrastructure needed to produce enough energy and chips to drive down the cost of compute and make it abundant,” it says. “In turn, this will create tens of thousands of skilled-trade jobs, boost local economies through spending and indirect job creation, and modernize our energy grid in the near term— ultimately supporting the kind of breakthroughs and innovations that drive lasting economic growth.”

That’ll work, provided AI ends up being as important to the long-term future of the world as OpenAI and all the other AI companies would have you believe.

OpenAI published its economic blueprint for America the morning the Biden White House announced sweeping new regulations of the industry. Biden’s new regulations will create a tiered list of countries that AI companies do business with. On tier one is the U.S. and 18 of its allies. Those countries are free of restrictions. China and Russia are on tier three and no AI companies can do business with them. The rest of the world is in tier 2. They can have a little AI, as a treat. But the White House will set caps.

“In its last days in office, the Biden Administration seeks to undermine America’s leadership with a 200+ page regulatory morass, drafted in secret and without proper legislative review,” NVIDIA said in a blog post about the new regulations. “This sweeping overreach would impose bureaucratic control over how America’s leading semiconductors, computers, systems and even software are designed and marketed globally.”

These seem to be the kinds of burdensome regulations that OpenAI railed against in its economic blueprint. But it’s also mostly focused on countries other than the U.S. Altman and OpenAI have long attempted to thread the needle of calling AI a revolutionary technology that needed to be unleashed while also claiming it’s dangerous and in need of serious regulation.

But the regulation-friendly Democrats are no longer in power. Everything could change in a week. Elon Musk is no fan of Altman and OpenAI and, for now, he’s close to the incoming president. It’ll be interesting to see how OpenAI’s calls for loose government regulation and widespread access to federal data are met by the new President and his Musk-backed administration.