The Only Way To Preserve America Is Through AI
(and the only way to manage it, is partnering with Startups & Big Tech)
When you analyze the paths the US and China have taken in the development of A.I. (artificial intelligence), it seems that the US has not been pushing the envelope. On the other hand, China has been pushing really hard. They’re focused on producing more Ph.D.s, and more papers, and more patents, and of course have simple more people to invest into the field. I am not saying that is better or worse, but it is clear that the US has not been putting the same resources into it. This is why the US has no choice at this point but to merge the interests of startups / big tech with big government, or be left in the dust (more on this later). First, let’s talk about China.
By many metrics China has surpassed the US in AI capabilities. We’re well aware of their facial recognition software, now implemented for years across their entire society. Less apparent but more powerful are the patterns that they’ve picked up from the plethora of data Chinese citizens have both willingly and innately provided the government’s model. China was engaging in data collection on a massive, granular scale long before most governments were aware of the implications. They have had more time to accumulate data on what works and what doesn’t than any other organized government in the world, and they’re applying this thinking across every channel they have access to, on a global level (think TikTok Meets Chinese Social Credit Algorithms). Still, by some metrics, there is still no question that the US is leading AI. The Chinese AI path is still more narrow — for example they are very focused on visual perception systems. While what they’ve accomplished is impressive, it is possible, even with tremendous effort, to succeed in a narrow domain and still lag behind in general intelligence. By most public-facing metrics, China is committed to AI superiority for peaceful purposes. They’re investing in it heavily across all their government agencies, and they’re combating the perception that they’re not playing by the rules with seemingly appropriate transparency initiatives — like opening up some of their A.I. labs to inspection by other leading researchers, where they are publishing the results of their research (keyword: some). Moreover, China is investing heavily in research around socially beneficial A.I., like education A.I., medical A.I., and environmental monitoring A.I.. And if you look at how much money goes into these areas relative to military or surveillance applications, it may seem obvious to the outside world that their orientation is toward peaceful purposes. I expect China will be far more transparent about their progress than the US has been or will be about 95% of their initiatives.
This is why, from my perspective, the US - in partnership with private corporations - has a responsibility to ensure that it stays ahead in AI development, both in the public-facing sectors like education and medicine, and the administrative sectors — ranging from elections to governance. As the face of democracy and capitalism, the US holds a special responsibility to guide the progress of AI in a way that benefits humanity. If we want to remain competitive in the global economy, we must make sure that America continues to push forward, making the key breakthroughs in key areas.
In my view, the government needs to do more to ensure that advances in AI are driven by American entrepreneurs and American research and that we remain the leaders in this critical area of technology. I’m hopeful that we can achieve this by:
- Investing more heavily in basic research and in ensuring that we have more of the world’s top talent working and building companies here. We need to take this golden opportunity to let in more highly-skilled tech workers, as well as more immigrants qualified and eager to study in relevant A.I. fields.
- Working closely with universities and companies on AI ethics and safety is a must. America is in desperate need of an A.I. Advisory board, which should be lead by people like Elon Musk, Sam Harris, and a number of top American entrepreneurs, luminaries and philosophers.
- Advancing new regulatory approaches — like innovation opportunity zones with tax incentives— that focus on managing the risks from rapid growth while preserving the potential benefits on the local, state, and national level.
The ultimate goal of the American government’s artificial intelligence initiatives should be to promote and protect America’s strategic interests and core values while maximizing global public trust and common security, while minimizing arms races and friction with allies and partners.
The United States should seek to maintain its technical advantage in AI while strengthening norms of responsible development among leading tech powers like China, Russia and potentially others. While I am hopeful that we can navigate this challenge, I am concerned that recent events show we are not ready to responsibly manage these challenges yet — not by a long shot.
The TL;DR Answer: Big Tech + Distributed Government
America’s AI winter has been discussed often over the past several years, the notion that AI has been overhyped and disappointed, that its success has been overstated. Certainly viewing from the lens of public government investment, America has not only been in a winter, but been absolutely dormant.
But I disagree with those who think it’s hopeless.
The news from the private sector American AI front is almost all good, and in fact sometimes embarrassingly good. The American tech industry is in the midst of a second artificial intelligence renaissance. Would it surprise you to know that almost the entirety of this article was written by Artificial Intelligence, based out of America? :wink:
The first renaissance was in the middle of the last decade, when many tech companies were convinced that AI’s big breakthrough was just one more funding round away. Now, a new wave of startups and tech giants alike are convinced that they can build intelligent systems to do all sorts of things: answer questions, produce podcasts, run companies, even drive cars. In the tech industry right now there’s a phrase that gets bandied about a lot: AI-first.
It refers to companies that have made artificial intelligence the central component of their product road map. In recent years, Google has been perhaps the most prominent company to adopt this strategy (though obviously not alone), with consumer efforts like its conversational assistant software for mobile devices and smart reply on Gmail, as well as pie-in-the-sky efforts in quantum computing (search: Time Crystals). But Google isn’t the only company pursuing an AI-first strategy; almost every tech company in America is investing in AI to help improve some facet of it’s business. Microsoft is leveraging its extensive research in machine learning to improve various Office apps, including partnering with Nvidia to create the world’s biggest language model; Amazon has been using machine learning techniques to make Alexa more useful for years; Apple has bought several small machine learning startups to better productize solutions for their “neural engine” chips (designed specifically for running machine learning algorithms). Even companies like Uber and DoorDash have applied machine learning to improve pickup times and route efficiency for its drivers.
And yet, something is still missing here, and that, my friends, is the partnership between corporations and government to advance the purposes of benevolent AI.
Imagine if the USPS ran as well as Gmail? What if Pinterest helped sort ballot initiatives? America needs to partner with corporations to help bootstrap AI in government, and in the American economy.
What will it take? Partnering with corporations to advance benevolent AI in government is all about taking what’s publicly funded — like open source code, big data sets, and the cloud (which is effectively free after you buy the hardware) — and applying it to government priorities. It does not mean swapping out systems wholesale, but rather applying the innovations of both new startups and established companies like Google, Amazon and Facebook to legacy systems like Defense, Health and Human Services, and State. Part of that will be putting pressure on those companies to open source code and data sets that can help ameliorate problems of governance.
Yes, the Pentagon needs large scale machine learning capabilities to help soldiers make better battlefield decisions — and yes, these investments have already been made. The partnerships that need to happen are the public-facing ones, that will make everyday life better in America, for every day people.
It’s time culture and people-based problems were more widely accessible to solve, as open source code, as well as application programming interfaces (API) through which other developers can access, use and build solutions for America and the globe’s most pressing problems. And it’s time the private sector was better incentivized and rewarded for doing so: what if Google entered into a partnership with HHS and got a 3% or 5% reduction in taxes owed if they helped HHS improve its data visualization capability?
Let’s also remember that some of these companies have very large R&D teams working on areas that can be put at the service of public institutions. Amazon has been working on image recognition tools for a long time. Google has been working on natural language processing tools from its inception. There are many places where these private sector innovations can be applied to public institutions, with public funding capitalizing those private sector innovations into profitable products for millions of consumers worldwide. In turn this could not only improve general quality of life, but lower taxes, reduce waste, and create jobs.
Our goal should not be to empower monopolies, but rather create a true free market in algorithms that help guide mankind. For governments to become more efficient through AI means we have to welcome technologies from many different sources so that we can benefit from competition between them. Competition builds better products at lower prices, and we need that competitive dynamic in American A.I.
Every company I know is tired of dealing with legacy technology; it makes their work harder; it slows them down; it makes them less agile; it limits their ability to hire people because they’re stuck dealing with old stuff; etcetera etcetera. But replacing existing technology is expensive (especially when you need an army of programmers to do it), which means that starting over is difficult for even big companies; let alone public institutions whose budgets are much smaller than these private sector giants.
By tapping into what’s already available in the private sector — like open source code — we can create new government capabilities quickly (and cheaply), allowing businesses to access them without investing millions or billions of dollars or spending years building their own versions. Within this partnership between government and corporations lies an enormous potential for accelerating the pace at which AI is applied within our economy, improving our lives while also creating jobs along the way.
Cal Chan is the CEO of Engaging, a company currently hiring designers, developers, dreamers and dooers.
