Over 100 government, academic, business, and technology leaders gathered at the White House recently to discuss Artificial Intelligence (AI). While there may have been intelligence in the room, very little was visible in the resulting document.
The summary of the meeting is here.
What is the government’s role in AI? While the government can help fund academic research, AI research in commercial spaces is moving forward at a blistering pace already. Many (including Bill Gates, Elon Musk, and the late Stephen Hawking) have worried that, going forward without any heed of consequences, AI could create computer systems that would threaten the species.
But from the summary, this meeting seemed to be about self-congratulatory cheerleading, devoid of any actual analysis or specifics — and with nothing about safety and regulation (except “getting rid of red tape”).
Let’s take a look at what was in there.
What is Artificial Intelligence? Basic question, no answer.
Here’s one definition: Artificial intelligence is the development of algorithms that mimic human thought.
This is vague. And it changes based on perception. The paradox is that once computers figure out how to do things, it becomes harder to think of them as artificial intelligence. Computers can beat humans at chess — but once that had happened, people stopped thinking if playing chess as artificial intelligence. Computers can recognize and interpret natural-language text. Is this AI, or just a clever algorithm?
This leads me to my own definition of AI, which is this:
Artificial Intelligence is the development of any computer system that mimics human capabilities, but in ways that its creators don’t fully understand.
This is much clearer, and of absolutely no value to research. Actual AI research concerns fields like machine learning, neural networks, natural-language processing, and autonomous systems (for example, autonomous vehicles). These things are much clearly defined, but lumping them together and calling them “AI” doesn’t really capture any specific meaning.
With this level of confusion, you would think that the first thing in the government’s report would be a definition of AI. Here’s what the report says AI is:
Yes. That’s right. There is no definition in the report. So the government just convened 100 people to discuss something, but we have no idea what.
A passel of platitudes
Despite the lack of a definition, the report includes plenty of broad statements about AI. My question is, what do they mean?
Here’s how it starts:
Artificial intelligence (AI) has tremendous potential to benefit the American people, and has already demonstrated immense value in enhancing our national security and growing our economy.
Then we get a couple of inspirational quotes from the President:
“We stand at the birth of a new millennium, ready to unlock the mysteries of space, to free the Earth from the miseries of disease, and to harness the energies, industries and technologies of tomorrow.” – President Donald J. Trump
“We’re on the verge of new technological revolutions that could improve virtually every aspect of our lives, create vast new wealth for American workers and families, and open up bold, new frontiers in science, medicine, and communication.” – President Donald J. Trump
Where is the AI in those quotes? Hmm. Of course, the executive summary is where the meat should be. Here it is, with jargon in italic and weasel words in bold, plus my commentary.
Key takeaways from breakout discussions included:
- Supporting the national AI R&D ecosystem. America is blessed with a unique R&D ecosystem that taps into the limitless bounds of American ingenuity. Attendees discussed our free market approach to scientific discovery that harnesses the combined strengths of government, industry, and academia and examined new ways to form stronger public-private partnerships to accelerate AI R&D.
Commentary: “Limitless bounds”? Weasel words are fine, but when they contradict each other it’s breathtaking.
- Developing the American workforce to take full advantage of the benefits of AI. AI and related technologies are creating new types of jobs and demand for new technical skills across industries. At the same time, many existing occupations will significantly change or become obsolete. Attendees discussed efforts to prepare America for the jobs of the future, from a renewed focus on STEM education throughout childhood and beyond, to technical apprenticeships, reskilling, and lifelong learning programs to better match America’s skills with the needs of industry.
Commentary: Platitudes. I was absent the day they made “reskilling” into a word.
- Removing barriers to AI innovation in the United States. Overly burdensome regulations do not stop innovation – they just move it overseas. Participants in this session addressed the importance of maintaining American leadership in AI and emerging technologies, and promoting AI R&D collaboration among America’s allies. Participants also raised the need to promote awareness of AI so that the public can better understand how these technologies work and how they can benefit our daily lives.
Commentary: I’ve seen no regulation of AI, unless you count regulations on self-driving cars. A careful inspection of the full document reveals no more specifics.
- Enabling high-impact, sector-specific applications of AI. Finally, attendees organized into industry-specific sessions to share the novel ways industry leaders are using AI technologies to empower the American workforce, grow their businesses, and better serve their customers.
Commentary: After reading four “key takeaways” I take nothing away. I learned that these people discussed AI but not what they said. This is a waste of pixels.
An inspiring speech
Michael Kratsios, Deputy Assistant to the President for Technology Policy, gave a speech at the event. Maybe the meaning is in there? You be the judge. Here are some excerpts, with my comments in brackets.
As artificial intelligence transforms everything from agriculture to manufacturing to transportation — the potential for AI remains breathtaking. [Maybe if you defined it first.]
But we cannot be passive. To realize the full potential of AI for the American people, it will require the combined efforts of industry, academia, and government. [I thought we were for free enterprise.]
Our Nation is the best place on earth to innovate.
According to industry studies, America not only has more AI startups than any other nation – nearly double our closest competitor – but America also has the most promising AI startups, with three-quarters of the world’s top 100. . . . [This is a hard point to prove. I don’t think you can measure AI innovation by sheer tonnage of startups.]
But I have also made clear that while America will always approach artificial intelligence prudently, we will not hamstring American potential on the international stage. [I feel safer already, don’t you?]
I want to focus on The White House approach to artificial intelligence and American industry—particularly how we can support the American worker, promote R&D, and remove barriers to innovation. [Nowhere in this report are any barriers actually cited.]
Artificial intelligence holds the promise of great benefits for American workers, with the potential to improve safety, increase productivity, and create new industries we can’t yet imagine. However, to a certain degree job displacement is inevitable. But we can’t sit idle, hoping eventually the market will sort it out. [Why not? Allowing the markets to sort things out is the traditional Republican approach to just about everything.]
Our free-market approach to scientific discovery harnesses the combined strength of government, industry, and academia, and uniquely positions us to leverage artificial intelligence for the betterment of our great Nation. [Now you sound more like a Republican.]
As we’re making great strides within the Administration, to the rest of America often the most significant action our government can take is to get out of the way.
Our Administration is not in the business of conquering imaginary beasts. We will not try to “solve” problems that don’t exist.
To the greatest degree possible, we will allow scientists and technologists to freely develop their next great inventions right here in the United States. [Still waiting to hear what these barriers actually are.]
We didn’t roll out the red tape before Edison turned on the first lightbulb.
We didn’t cut the lines before Alexander Graham bell made the first telephone call.
We didn’t regulate flight before the Wright Brothers took off at Kitty Hawk. [Regulation trails innovation. ‘Twas ever thus. What’s different about AI?]
We can only succeed if the American people succeed— only if they have jobs they can be proud of, only if they have neighborhoods and cities on the rise, only if their lives and relationships are enriched, not estranged, by the technology we invent.
I have every reason to believe we can accomplish this. Generation after generation, American innovation has benefited our people and the entire world. [Meaningless platitudes.]
Today, with so many of the mysteries of quantum computing, autonomous systems, and machine learning yet to be discovered, we can take hold of the future and make it our own. [This, finally, is a sentence that actually describes a set of AI technologies. Except that calling quantum computing “AI” isn’t quite right. But since we never defined AI, I guess you can put anything in there that you want.]
The collision of the two forms of Trump Administration bullshit
There are two forms of bullshit in the Trump administration.
The first is statements by Trump himself that are simple, declarative, and bear little relationship to the truth. This is the bluster at the top, and we’ve seen a lot of it, with over 2000 false statements in the first year alone.
The second is traditional government bullshit: vague platitudinous statements filled with jargon that actually have no content.
This document includes both. It’s filled with both the simple, direct, and irrelevant statements (“We didn’t roll out the red tape before Edison turned on the first lightbulb.”) and the usual meaningless platitudes (“Our free-market approach to scientific discovery harnesses the combined strength of government, industry, and academia, and uniquely positions us to leverage artificial intelligence for the betterment of our great Nation.”)
AI is filled with challenges and opportunities. I expect America to continue to develop it, as will China and other countries. The government will fund some of it. And, like any other industry, it needs commonsense regulation.
After 100 people got together in Washington to talk about it, I see no evidence that anything has changed.
5 responses to “The Trump Administration’s take on Artificial Intelligence (AI)”
I have read extensively about “AI” which is more accurately described as machine learning.
The best thing the public sector could do in relation to it is as little as possible from the legislative and regulatory perspective. As an IT consumer, government should use it to the greatest extent possible.
I disagree with your definition of AI. Machine learning is a subset. Is a car that is driving autonomously “AI”? Is Alexa “AI”? I don’t think those qualify as just machine learning.
I completely agree with you about regulation.
Yes. Machine learning is a subset of AI.
“Our free-market approach to scientific discovery harnesses the combined strength of government, industry, and academia, and uniquely positions us to leverage artificial intelligence for the betterment of our great Nation.”
“Artificial Intelligence” generally refers to research on computer algorithms that is inspired by the way that humans think or behave. Once we understand an algorithm we often stop calling it “artificial intelligence” because it is no longer research. However some algorithms (e.g. neural nets) that came out of “AI research” retain that title.
Interest in artificial intelligence tends to peak every decade or two with useful breakthroughs but excessive optimism leading to disappointment. This time might be different, of course.
I am personally very skeptical about the near term hopes for self-driving cars.
The self-driving car is precisely the kind of problem where you think you are 95% of the way solved when you are actually more like 10% of the way solved. This is because of a combination of factors: (a) the domain is incredibly unstructured (roadways, driver behaviors, pedestrian behaviors), and (b) it’s a matter of life and death. An additional issue (c ) the potential for an automobile to be hacked, or entire fleet of automobile system intercommunication to be hacked. To me it looks like a fascinating research problem with a long future 😉
I won’t make the obvious remark about why this administration needs artificial intelligence, although I’m sure many of your readers thought the same thing when they read your headline. But I do wonder how all these platitudes about new technology and “reskilling” resonate with all those voters in the Rust Belt who are waiting for Trump to open their factories again.