Maybe AI Doesn't Read Blueprints

Gosh, who knew that today would be an AI day, with at least three major announcements about “blueprints” for its development going forward? Of course, these days every day is an AI day; trying to take in all AI-related news can be overwhelming. But before some other AI news drowns them out, I wanted to at least outline today’s announcements.

The Biden Administration thinks it has an AI blueprint. Credit: AI.gov

The three I’m referring to are the Biden Administration’s Interim Final Rule on Artificial Intelligence Diffusion, OpenAI’s Economic Blueprint, and the UK’s AI-driven Plan for Change.  

The Biden Administration’s rules aim to preserve America’s lead in AI, stating: “it is essential that we do not offshore this critical technology and that the world’s AI runs on American rails.” It establishes who advanced chips can be sold to and how they can be used in other countries, with no restrictions on 18 key allies and partners.

It also sets limits on model weights for AI models, seeking to constrain non-preferred entities’ ability to train advanced AI models.

“The U.S. leads the world in AI now, both AI development and AI chip design, and it’s critical that we keep it that way,” Commerce Secretary Gina Raimondo said in a briefing with reporters ahead of Monday’s announcement

Not everyone is happy. The Information Technology & Innovation Foundation blasted the rule, claiming it would hamper America’s competitiveness.  Vice President Daniel Castro warned: “By pressuring other nations to choose between the United States and China, the administration risks alienating key partners and inadvertently strengthening China’s position in the global AI ecosystem.”

Similarly, Nvidia, which makes most of those advanced AI chips, expressed its opposition in a statement from Ned Finkle, vice president of government affairs, claiming the rule “threatens to derail innovation and economic growth worldwide.”  He explicitly contrasts how the first Trump Administration (and, one assumes, the next Trump Administration) sought to foster “an environment where U.S. industry could compete and win on merit without compromising national security.”  

Not to be outdone, Ken Glueck, Executive Vice President, Oracle, says the rule “will go down as one of the most destructive to ever hit the U.S. technology industry,” and “we are likely handing most of the global AI and GPU market to our Chinese competitors.”

It will be interesting to see what the Trump Administration does with the Rule.

Meanwhile, OpenAI’s economic blueprint believes “America needs to act now to maximize the technology’s possibilities while minimizing its harms…to ensure that AI’s benefits are shared responsibly and equitably.” Its goals are to:

• Continue the country’s global leadership in innovation while protecting national security
• Make sure we get it right in AI access and benefits from the start
• Maximize the economic opportunity of AI for communities across the country.

It sees “infrastructure as destiny,” with investment in AI infrastructure “an unmissable opportunity to catalyze a reindustrialization of the US.” It wants to ensure that “an estimated $175 billion sitting in global funds awaiting investment in AI projects” get invested here rather than in China.

OpenAI does want “common-sense rules” that promote “free and fair competition” while allowing “developers and users to work with and direct our tools as they see fit” under those rules. And, of course, all this while “Preventing government use of AI tools to amass power and control their citizens, or to threaten or coerce other states.” It particularly wants to avoid a “patchwork of state-by-state regulations”

The company is planning an event in Washington D.C. on January 30 with CEO Sam Altman “to preview the state of AI advancement and how it can drive economic growth.”  I’ll bet Mr. Altman is hoping he gets plenty of Trump Administration officials, although probably not Elon Musk.

Credit: OpenAI

Last but not least, UK Prime Minister Keir Starmer has endorsed an ambitious set of AI recommendations, wanting to turbocharge the economy by turning the UK into an AI superpower. Mr. Starmer vowed:

But the AI industry needs a government that is on their side, one that won’t sit back and let opportunities slip through its fingers. And in a world of fierce competition, we cannot stand by. We must move fast and take action to win the global race.

Our plan will make Britain the world leader. It will give the industry the foundation it needs and will turbocharge the Plan for Change. That means more jobs and investment in the UK, more money in people’s pockets, and transformed public services.

There are three key elements:

First, “laying the foundations for AI to flourish in the UK,” including AI Economic Growth Zones and  a new supercomputer.

Second, “boosting adoption across public and private sectors,” such as through a new digital government center that “will revolutionise how AI is used in the public sector to improve citizens lives and make government more efficient.”

Third, “keeping us ahead of the pack,” with a new team that “will use the heft of the state to make the UK the best place for business.”

It will do so while also charting its own course on regulation. "I know there are different approaches (to AI regulation) around the world but we are now in control of our regulatory regime so we will go our own way on this," the PM said. "We will test and understand AI before we regulate it to make sure that when we do it, it's proportionate and grounded."

Credit: Gov.UK

Chris Lehane, Chief Global Affairs Officer at OpenAI, praised the plan: “The government’s AI action plan - led by the Prime Minister and Secretary Peter Kyle - recognises where AI development is headed and sets the UK on the right path to benefit from its growth:”

All nice words, but lots left unsaid. As Gaia Marcus of the Ada Lovelace Institute pointed out: "Just as the government is investing heavily in realising the opportunities presented by AI, it must also invest in responding to AI’s negative impacts now and in the future."

-----------

These things are true: AI is going to play a major role in the world economy, and to be a superpower, a country will have to be an AI superpower. To be an AI superpower, a country has to have the best AI infrastructure, including chips and data centers. AI is equally capable of positive impacts as well as negative impacts, and some regulation is needed to mitigate the latter. Lastly, regulation is going to lag innovation -- and AI will drive innovation at rates we haven’t seen before.

I envy the people working on AI innovation, but I don’t envy those trying to figure out how to best regulate it.