Search Episodes
Listen, Share, & Support
Listen to the latest episode
Subscribe via iTunes
Subscribe via RSS
Become a fan
Follow on Twitter

Support Us:

Please consider making a donation to help make this podcast possible. Any contribution, great or small, helps tremendously!

Subscribe to E-Mail Updates

Related Readings
  • Answers for Aristotle: How Science and Philosophy Can Lead Us to A More Meaningful Life
    Answers for Aristotle: How Science and Philosophy Can Lead Us to A More Meaningful Life
    by Massimo Pigliucci
  • Nonsense on Stilts: How to Tell Science from Bunk
    Nonsense on Stilts: How to Tell Science from Bunk
    by Massimo Pigliucci
  • Denying Evolution: Creationism, Scientism, and the Nature of Science
    Denying Evolution: Creationism, Scientism, and the Nature of Science
    by Massimo Pigliucci

RS 231 - Helen Toner on "Misconceptions about China and artificial intelligence"

Release date: April 15th, 2019

Helen Toner

Helen Toner, the director of strategy at Georgetown's Center for Security and Emerging Technology (CSET), shares her observations from the last few years of talking with AI scientists and policymakers in the US and China. Helen and Julia discuss, among other things:

  • How do the views of Chinese and American AI scientists differ?
  • How is media coverage of China misleading?
  • Why the notion of an "AI arms race" is flawed
  • Why measures of China's AI capabilities are overstated
  • Reasons for optimism and pessimism about international cooperation over AI


"Age of Ambition: Chasing Fortune, Truth, and Faith in the New China" by Evan Osnos

"The Beautiful Country and the Middle Kingdom: America and China, 1776 to the Present" by John Pomfret

Center for Security and Emerging Technology

A recent article co-authored by Helen, in Foreign Affairs: "Beyond the AI Arms Race: America, China, and the Dangers of Zero-Sum Thinking"

Follow Helen Toner on Twitter

Edited by Brent Silk

Music by Miracles of Modern Science

Full Transcripts 

Reader Comments (8)

Wow, that's what I was searching for, what a material!present here at this website, thanks admin of this web site.
Really great podcast. This issue definitely merits more public attention.

Good Foreign Affairs article. Innovation and Development will indeed have more impact than either (1) massive amounts of "big" data, (2) speculative investment, or (3) the combination of (1) and (2). Progress will come from primary research and new discoveries.

The Western News Media definitely did misrepresent and over dramatize the China Social Credit Score System (no real surprise though).

China aggressively manipulates its GDP numbers in order to meet the demands of government leadership. China also has entire vacant cities built in order to increase construction output.

One might seriously doubt that Chinese citizens actually have more acceptance of their government having their private data than corporations having that data. That seems like a statistic that the communist government itself either directly produced, or else procured the services of some news or analysis organization to produce.

Regulation of AI will probably ultimately come from the Federal Trade Commission (FTC) and other consumer protection agencies. Real rapid advances in AI applications will probably originate with the commercial sector. Most governments lack the adaptability, insight, and responsiveness to consumer demand necessary to create highly marketable products that can grow the industry.
April 16, 2019 | Unregistered CommenterJameson
Fundamental machine learning algorithms haven't evolved much in decades, but computing power and volume of data has increased. The U.S. collects a lot of data, but it's stovepiped. China's data may be more centralized.

Google invests a lot in AI and owes a lot to DARPA, but then refuses to develop AI for the U.S. military while opening an AI center in China. Microsoft and Amazon have been less treacherous so far, and there are plenty of other companies and labs that are happy to do U.S. defense work, as you can see in the DARPA AI Colloquium.

Of course after Google made its TensorFlow library open-source, anyone can use it including China and North Korea for that matter. And anyone can read publications and patents. To stay ahead of competitors, you have to keep information secret, but the U.S. has had a hard time doing that.

Autonomy is an important part of everyone's defense strategy, and you can compare who's ahead in developing and deploying unmanned systems. China is reportedly throwing a lot of money at it. Don't underestimate them the way we underestimated North Korea's nuclear program.
April 18, 2019 | Unregistered CommenterMax
Je vous aiderai à faire ce projet en cas de succès, laissez-moi l'essayer, merci
April 23, 2019 | Unregistered Commenterslither io
Why is it called it a "social credit score" if it does not function like one?

The argument that it is not "a single unified number ... that rules every aspect of your life" is fallacious - a bit of straw man and a bit of false dichotomy. The claim could be literally true, yet meaningless (e.g., there could be multiple numbers - some of them could be based on things that have nothing to do with adherence to law or financial creditworthiness). Not to mention that adherence to law in some countries means not expressing dissent. Social credit for conformity and obedience.
April 29, 2019 | Unregistered CommenterTimeZ
The nuanced way that both parties discussed, excepted, caveated, uncertainified, and bias-acknowledged, gave me this ... sigh of relief. As if recognizing my tribe. They earnestly strove to understand better.

This podcast is more than an oasis of rationality. It's kind of an oasis of enlightened irrationality.
May 9, 2019 | Unregistered CommenterBob Stein
Far from ruining its economy in the 50s and 60s, China–under ferocious embargoes against food, finance, technology and international participation, outgrew America's economy by 300%. National income grew five-fold over the 25-year period 1952-78, increasing from 60 billion to over 300 billion yuan, with industry accounting for most of the growth. On a per capita basis, the index of national income (at constant prices) increased from 100 in 1949 (and 160 in 1952) to 217 in 1957 and 440 in 1978.

Over the last two decades of the Maoist era, from 1957 to 1975, China's national income increased by 63 percent on a per capita basis during this period of rapid population growth, more than doubling overall and the basic foundations for modern industrialism were laid and outpacing every other development takeoff in history.

- In Germany the rate of economic growth 1880-1914 was 33 percent per decade.
- In Japan from 1874-1929 the rate was 43 percent.
- The Soviet Union over the period 1928-58 the rate was 54 percent.
- In China over the years 1952-72 the decadal rate was 64 percent.

Bear in mind that, save for limited Soviet aid in the 1950s, repaid in full and with interest by 1966, Mao's industrialization proceeded without benefit of foreign loans or investments–under punitive embargoes the entire 25 years–yet Mao was unique among developing country leaders in being able to claim an economy burdened by neither foreign debt nor internal inflation.

Without Mao's industrial revolution, the economic reformers of the post-Mao era would have had little to reform because the higher yields obtained on individual family farms during later years would not have been possible without the vast irrigation and flood-control projects–dams, irrigation works and river dikes–constructed by collectivized peasants in the 1950s and 1960s.

About today, there is nothing ‘true’ or ‘false’ about GDP growth figures because countries can define GDP as it suits them. China chooses and weights components of her gross domestic product differently than America, so it’s hard to make apples-to-apples comparisons. But let’s try.

There is, for example, ten times more private real estate construction going on in China right now than in the US but real estate contributes only six percent to China’s GDP while A American construction represents fifteen percent of hers.
That’s because China looks at GDP from the point of view of a social necessity, a utility everybody needs (socialism) while America sees GDP from the point of view of a bunch of individual, greedy, amoral consumers (capitalism).

A American economists base their calculations on the imputed cost of renting every owner-occupied home in the country. That’s how a bunch of individual, greed-crazed, amoral consumers would see it, so that’s how they count it and, if the US Government switched to counting housing as a utility, like the Chinese do, there’d be a revolution.
Chinese economists count urban housing at its purchase price amortized over fifty years and totally ignore a billion square meters of new rural housing every year. To them there’s no tradeable ‘value’ in it and there should not be. Since it’s a necessity, only its use value should be counted and everything should be geared away from profits and towards getting everyone into a decent home. This approach produces more democratic results than America’s:

Of course, if the Chinese Government switched to counting housing as renting every owner-occupied home in the country, like the A Americans do, there’d be a revolution.
lso, because Communists don’t have servants, China excludes the domestic product of fifty-million live-in nurses, millions of babysitters, cleaners, cooks and tutors and women selling jianbing snacks curbside in Beijing who clear $600 a day. They don’t count shadow banks either, or a million Uber/Didi drivers, or a hundred million vendors on Taobao, the six hundred billion dollar e-market. nd outside Beijing and Shanghai all transactions–including automobile and real estate purchases–are conducted in cash by people who’ve been avoiding taxes for two thousand years.

Bottom line? China’s GDP is much bigger than its GDP figures disclose and, therefore, so is China’s GDP growth. Don’t confuse GDP growth rates–acceleration–with actual growth: they are distinctly different. In 2006, China’s GDP was $7.6 trillion and the economy accelerated by 12 percent (compared to the previous year) or $761 billion. In 2016 GDP was $21.1 trillion and the economy accelerated by 6.7%, it grew by $1.3 trillion, twice as much as in 2006.

Taking methodological differences into account, Washington’s Center for Strategic and International Studies calculated that China's economy is fifteen percent larger than official figures and the Peterson Institute of International Economics went further, estimating[1] that it’s twenty-seven percent bigger and on track for $29 trillion in 2020: fifty percent larger than America’s.

So China is (quite legally and ethically and publicly) understating the size of her economy by, let’s say, 30%, which means that China has been systematically understating her GDP growth for a long time…by our standards.
But should we force her to change to our method of calculating GDP and, if so, why?
[1] Is China lready Number One? New GDP Estimates. rvind Subramanian (PIIE). January 13, 2011 5:15 PM
May 20, 2019 | Unregistered CommenterGodfree Roberts
I appreciate the interview. A piece of friendly advice, given with sincere empathetic kindness, is that our use of the term "sort of", as well as other fillers such as "kind of" and "like", can make our communication less smooth and, I suspect, make others perceive us as less competent.
June 15, 2019 | Unregistered CommenterMC

PostPost a New Comment

Enter your information below to add a new comment.

My response is on my own website »
Author Email (optional):
Author URL (optional):
All HTML will be escaped. Hyperlinks will be created for URLs automatically.