US, China Take the Lead in Race For AI: UN (reuters.com)

China and the United States are ahead of the global competition to dominate artificial intelligence (AI), according to a study by the U.N. World Intellectual Property Organization (WIPO) published on Thursday. From a report: The study found U.S. tech giant IBM had by far the biggest AI patent portfolio, with 8,920 patents, ahead of Microsoft with 5,930 and a group of mainly Japanese tech conglomerates. China accounted for 17 of the top 20 academic institutions involved in patenting AI and was particularly strong in the fast growing area of "deep learning" - a machine-learning technique that includes speech recognition systems.

"The U.S. and China obviously have stolen a lead. They're out in front in this area, in terms of numbers of applications, and in scientific publications," WIPO Director-General Francis Gurry told a news conference. U.S. President Donald Trump has accused China of stealing American innovations and technology and has slapped trade tariffs on $234 billion of Chinese goods to punish Beijing.

Comments

Wait before you draw conclusions

[Opportunist]

They hold the most patents. That's all. We're still a far cry from answering the all important question: What does "AI" mean anyway?

Re:Wait before you draw conclusions

[DigiShaman]

China is throwing human resources at all levels. Please see their hyper-sonic warhead and rail-gun development. In fact, the dirty little secret is that despite the rivalry between US and China, the real actors to watch out for is how both India and Russia respond. Those two won't take this Chinese development laying still.

Re:Wait before you draw conclusions

[jellomizer]

Well there is the public Idea of the Commander Data, or Hal 9000 type of AI. What modern AI is a computer used to solve complex problems, and be adaptable to come up with a unique solution.
What has changed over the past few decades, is that computing power has gotten to a level that affordable computers can now perform these calculations. Modern AI is still kinda stupid, but it is more rigorous.
Say given 50 years of crop data, with 50 years of weather data. Correlating, Simulating and Trending the data use to be too expensive, now it can be ran to show when it would be a good time to irrigate your crops and when to wait for rain, for your area. So either it will take a team of experts to try to figure this out over years. Or an AI that will just run for a few days and give you data that may be good enough.

Re:Wait before you draw conclusions

[fat man's underwear]

I think it means "Annual Inflation" of capabilities and promises.
Since our society seems to be based on the flow of money, and all our physical needs have been met a long time ago, we have to find new ways to get that money moving.
Fads, fashions, trends... aren't just for clothes!

Re:Wait before you draw conclusions

[phayes]

Does anyone really think that because IBM/Microsoft/... has patented something that that will stop anyone in _China_ from using/infringing it?

Re:Wait before you draw conclusions

[gweihir]

We're still a far cry from answering the all important question: What does "AI" mean anyway?

That one is solved: "AI" = marketing hype term meaning "automation".

Re:Wait before you draw conclusions

[ranton]

They hold the most patents. That's all. We're still a far cry from answering the all important question: What does "AI" mean anyway?

Very few people are wasting their time on that question. They are finding real world uses for modern AI techniques, and expanding on that toolbox of techniques, not wasting their time on pedantic argument like whether or not this will help lead to strong/general AI.

It is still one of many important questions in the field of AI, but it is irrelevant to nearly all of the many uses AI has for us right now.

Re:Wait before you draw conclusions

[Opportunist]

How can you use something that you can't even define? Flumbo is good for you, Flumbo will make you big, strong and wow! Wtf?

Re:Wait before you draw conclusions

[ranton]

How can you use something that you can't even define? Flumbo is good for you, Flumbo will make you big, strong and wow! Wtf?

There are plenty of definitions within the AI field. Machine learning, deep neural nets, decision trees, etc. A set of terminology is obviously necessary to do anything in the field.

But worrying about a general definition of AI is kind of like trying to define what consciousness truly is. Still a useful endeavor, but not that useful to the doctor trying to revive an unconscious patient.

Re:Wait before you draw conclusions

[K. S. Kyosuke]

and all our physical needs have been met a long time ago

I'm sure the ultrarich disagree; you still need to get more productive for them to grift some more money from you. So there's still a lot of productivity pressure.

Re:Wait before you draw conclusions

[hackingbear]

Does anyone really think that because IBM/Microsoft/... has patented something that that will stop anyone in _China_ from using/infringing it?

It will get China to stock up their weapons to fight off American patent trolls.

Re:Wait before you draw conclusions

[religionofpeas]

How does a bird define flying ?

Re:Wait before you draw conclusions

[phayes]

That mostly works outside China. My point being that Chinese spies steal tech and then hand it over to chinese companies to use without any care in it being patented or not.

Re:Wait before you draw conclusions

[hackingbear]

1. patent infringements are not the same as IP theft by definition;
2. patent infringements happen all the time in the US, check the East Texas court filings.
3. and patents are also enforced in China. in fact, it is known that foreign patent holders get favorable ruling over there.
4. at the same time, patent trollings also happen all the time in the US, also check the East Texas court filings.
5. The US also commit IP theft even nowaday, in addition to its historic dirt.

Re:Wait before you draw conclusions

[AHuxley]

AI is the magic word that gets funding from the mil and gov in most advanced nations.
No AI winter just yet https://en.wikipedia.org/wiki/... .
Until then its AI everything and budgets get approved.

Re:Wait before you draw conclusions

[phayes]

China also has laws guaranteeing human rights but Tien an Men and their well known "selective enforcement" of their laws _for_ those in power and _against_ dissidents and barbarians (non-Han) prove, those "laws" are of vanishingly small use.

East Texas was a zealous enforcer of _bad_ patents. Thankfully the USSC's ruling in Oil States Energy Services LLC vs Green's Energy Group and other recent decisions has nerfed most of those going forward.

China enforces those laws that those in power chose to enforce against those it wants to and ignores the rest.

Now before you go all whataboutist with the US internment of Japanese US citizens and Kent State and Trump and whatever, The U.S. either made those mistakes in the past or they are being countered by the rule of law -- something which still exists here but not in china.

So sorry for you that _another_ chinese citizen was picked up for spying on Apple's autonomous vehicle with the intent to sell it to the chinese company he intended to work for.

Re:Wait before you draw conclusions

[hackingbear]

The US also has gitmo that locks up without trial people the government does not like.

Whataboutism? Ho ho ho. This article is about AI race and it is used promptly to attack Chinese using IP issue. What about that?

And then what about hypocrisy? what about double standard?

Re:Wait before you draw conclusions

[phayes]

Gitmo isn't imprisoning "people that the USG does not like". Gitmo is imprisoning extremist terrorists that were foreign combatants that believed in unrestrained terrorist acts against almost every government & people on earth. Even so, almost all of those imprisoned in Gitmo have been released back into their own government's care -- except for the 40 hardest core & those who the USG believes would be summarily executed.

TFA is about using patents as a means of comparing research which is meaningless when China only enforces "their" patents and ignore everyone else's. That's not anti-Chinese, it's anti-oppressive chinese government.

Speaking of hypocrisy & double standards, in which prisons would you be disappeared to were you to be Uigur or Tibetan or even a HAN student and expressing your opposition to one-party rule and wanting to vote for open elections? Yeah, that's right, in those Chinese prisons that harvest organs off of the imprisoned. Yeah for you it's better to crush your opposition under tank treads the Tien An Men way and disappear them rather than document them the way the U.S. has.

You're an apologist for the dictators and oppressors.
Bye troll-boy.

Re:And why, may I ask

[olsmeister]

How will we know if we're winning or losing if there is nobody keeping score?

Re:And why, may I ask

[jellomizer]

The UN goal is to help facilitate peace. Scarcity causes conflict and wars. Because the Orphans, Homeless refugees are this way because of this.
To help manage Scarcity we need to know who may have more and less of something, where the UN can help facilitate helping those with less of something to trade with countries you may have more.

While Water, Food, and Raw Materials are often big on the list, Intellectual Property and Technology are important as well. Knowing that the United States and China are leaders in such technology, a smaller country who may need such services may go to the UN to help them find a partner to work with them on getting such services. Where payments and negotiation needs to be taken place.

Now for some countries they really don't like China or the United States. However if given both they may choose one over the other.

That is why the UN is figuring out this stuff. Because if some unknown country who doesn't have access to modern AI research, say in crop management, and they are suffering from a food shortage, while their neighbors who have access to this technology and are doing better... Just may be enough to cause a war.

So..

[fluffernutter]

So the Chinese and the US will be out of work first.

Re:So..

[fluffernutter]

Oh I mean, they can show us the new really wonderful jobs that the new AI economy create.

UN money for war program

[Anonymous Coward]

The same UN that facilitated illegal oil for money trades with Iraq under Saddam to make Kofi Annan's son wealthy. Which led directly to the reasons for the US invasion of Iraq.
Details

That UN that supports corruption for their leadership's personal gain even if it causes war and mass causalities? The fact that they haven't done anything about that incident shows the UN isn't about keeping peace, they are about collecting money and power and not above doing so unethically.

Re:UN money for war program

[jellomizer]

I said it was its goal. Not necessarily that they are good at it, or always succeed.

If it's a race, what is the finish line?

[DavenH]

They give the feeling that we're developing the field toward some magic ignition point, then it's going to be worthy of calling "true AI." AI is here already. Maybe they mean AGI? OTOH they probably don't know what they mean.

Re:If it's a race, what is the finish line?

[gweihir]

AI used to mean AGI, until some people found they desperately wanted to give a false impression and invented "AGI", so they could continue to call things that have absolutely no insight or intelligence "AI".

The utterly dishonest claim that AI will eventually become AGI is just part of the scam being perpetrated here. There is no indication AGI will ever be possible and quite a few that it may not be.

Re:If it's a race, what is the finish line?

[DavenH]

That "AI used to mean AGI" is slightly true. Everything else you claimed is rubbish.

Re:If it's a race, what is the finish line?

[K. S. Kyosuke]

AI used to mean AGI

I thought it used to mean the field that tries to build systems that perform such tasks that if they were done by a human, we'd take it as a sign of his intelligence, but that's just me (and Minsky).

Re:If it's a race, what is the finish line?

[religionofpeas]

Military superiority is one of the goals.

Re:If it's a race, what is the finish line?

[gweihir]

If you think you can implement insight in computers today, then you are utterly without insight. Also, that there is absolutely no indication that AGI is possible is just the current scientific state-of-the-art. You seem to confuse SF literature with Science.

Re:If it's a race, what is the finish line?

[gweihir]

I thought it used to mean the field that tries to build systems that perform such tasks that if they were done by a human, we'd take it as a sign of his intelligence, but that's just me (and Minsky).

Well, Minsky was perhaps the most clueless idiot or alternatively the most dishonest ass with regards to what he said about AI. I am really glad that has finally stopped. It was an utter disgrace to the profession and is one of the sources of completely unrealistic expectations in non-experts.

What we are learning is that some things can be faked or actually done without intelligence. Chess and Go are current examples. These machines are not intelligent, they just managed to scale to a level where somewhat refined brute-force can beat a not-too-well prepared human expert player. (The machines had hundreds of games of their human opponents, the humans had none of the machine. That is called a "stunt" and has no scientific value at all.) Also, things like driving clearly do not require intelligence, even though many humans like to think otherwise. Bad translations also seem to not require intelligence, as long as you stick to a predefined set of topics (which is impressively large these days, granted). And, of course, the average human is not very smart and rarely acts insight-driven. That does not change the fact that machines are completely incapable of insight and that there is not even a credible theory these days how they could gain that. This means typically > 50 years or impossible. And no, IT is not faster developing than other tech. Look at the current mess with all that insecure crap and then compare that to the steam-engine: About 50 years to general adoption, but unsafe as hell and in no way a mature technology.That is where we are with computers today. The thing is that technology maturation is driven by humans and humans are generally slow and dumb.

Re:If it's a race, what is the finish line?

[K. S. Kyosuke]

What we are learning is that some things can be faked or actually done without intelligence. Chess and Go are current examples. These machines are not intelligent, they just managed to scale to a level where somewhat refined brute-force can beat a not-too-well prepared human expert player

So you're of the "if it can be done by a machine, it's not intelligence" crowd? I'm pretty sure that we'll eventually discover that there's no such thing as intelligence in the first place one day this way.

Re:If it's a race, what is the finish line?

[gweihir]

What we are learning is that some things can be faked or actually done without intelligence. Chess and Go are current examples. These machines are not intelligent, they just managed to scale to a level where somewhat refined brute-force can beat a not-too-well prepared human expert player

So you're of the "if it can be done by a machine, it's not intelligence" crowd? I'm pretty sure that we'll eventually discover that there's no such thing as intelligence in the first place one day this way.

Nope. Just what _these_ machines (or any other today) do is not intelligence. There is no element of insight. That may eventually change, but so far nobody has a clue how that could be achieved. If you think the element of insight is not needed for intelligence, then you are lacking in same.

Re:If it's a race, what is the finish line?

[K. S. Kyosuke]

I'm not sure why the term "insight" should be any more specific or objectively defined than "intelligence".

Re:If it's a race, what is the finish line?

[gweihir]

Well, the problem here is on your side. And it does not look like it can be fixed. Sorry.

Re:If it's a race, what is the finish line?

[K. S. Kyosuke]

So your saying that philosophy of mind is a solved problem? Interesting! I must have missed that.

Re:If it's a race, what is the finish line?

[K. S. Kyosuke]

It's not even slightly true.

Re:If it's a race, what is the finish line?

[RespekMyAthorati]

There is no indication AGI will ever be possible and quite a few that it may not be.

Ah, once again gweihir makes that same claim (as he has done dozens of times before)
without providing the slightest shred of evidence to support it.

I guess he thinks that if he spreads enough bullshit around, people will believe him.

Re:If it's a race, what is the finish line?

[RespekMyAthorati]

So you're of the "if it can be done by a machine, it's not intelligence" crowd?

That's exactly what he is, but he will dishonestly deny it.

Re:If it's a race, what is the finish line?

[RespekMyAthorati]

Well, the problem here is on your side. And it does not look like it can be fixed. Sorry.

In other words:
How dare you deny my philosophical insights! I am gweihir, the all-powerful!

Re:If it's a race, what is the finish line?

[gweihir]

Nope. There are just some (few) people that get how utterly limited and primitive modern computer tech actually is, and many, many more that ascribe magical powers to it and think it can do anything. The second class has a problem with reality perception that, I think, cannot be fixed. It seems to apply here.

Re:If it's a race, what is the finish line?

[gweihir]

I am saying it is _unsolved_ and there are pretty strong hints that we are nowhere close to even a rough solution. On the other hand, we do have a pretty strong theory of computing and that one gives us severe theoretical and practical limits as to what can be done with computing. Implementing consciousness is not possible with computers, and that is just a hard fact. All you could ever do is fake it. Of you think that is enough, then you have not understood what a p-zombie is and why it cannot be all that exists. AGI may well turn out to be impossible as well, but there is still some hope.

Why is it that people always try to twist the words of others when they hear something they do not like?

Unsuitable metrics

[gweihir]

Patent-count is not a measure of leadership. It is a measure of ego, greed and lack of ethics. In that, China may indeed have overtaken the US.

Re:Unsuitable metrics

[ranton]

Patent-count is not a measure of leadership. It is a measure of ego, greed and lack of ethics. In that, China may indeed have overtaken the US.

It is basically a measure of investment. You aren't going to invest significant money in research without protecting that investment with patents. Not unless you just consider the investment to be a donation to the common good.

That said, it is a very imprecise measurement because not all investment leads to significant returns. Coca-Cola investment millions in creating "New Coke" decades ago, and all of that investment was lost. Any patent on the New-Coke formula is essentially worthless. But taken in the aggregate, such as the patent portfolio of an entire country, it is probably a very good proxy of total R&D spend and the return on that investment in the industry of AI.

Re:Unsuitable metrics

[PolygamousRanchKid]

Patent-count is not a measure of leadership.

Well, IBM AI did defeat the world chess champion, Gary Kasparov. And it defeated all humans at Jeopardy, as well.

I dunno if the Chinese play Jeopardy, but maybe we could have a match of IBM AI vs. Chinese AI for the World Championship . . . ?

Does Watson speak Chinese . . . ?

On the other hand, I seem to remember some new a while back that IBM Watson AI Health wasn't doing so well.

Maybe patents != products . . . ?

Re:Unsuitable metrics

[hackingbear]

Nowaday, patent portfolio is your war chest. If you don't have a lot of patents, you are at risk of being sued out of existence by US patent trolls. Given the US wants to fight this "intellectual property" war, i.e. patent trolling, it is only natural for Chinese companies to stock up their weapons.

Re:Unsuitable metrics

[Guybrush_T]

It's a very bad way to measure investment.

The companies pushing AI right now are mostly Google, Facebook, Amazon, Apple ; their Chinese counter parts : Baidu, Alibaba, Tencent ; not to forget some Japanese companies as well like Preferred Networks or Sony. Universities of the same countries also play a part of it.

Those companies and universities do not care much about patents, at least much less than patent factories like IBM.

Note I'm not saying IBM is not investing on AI, just that the patents do not reflect the key papers that actually make science move forward.

Re:Unsuitable metrics

[Guybrush_T]

That was decades ago. And who defeated humans at Go ? Google. Ask yourself why Google is not the #1 patent holder on AI. Because patents are a bad way of measuring investment. Some companies are patent factories, some don't.

Re:Unsuitable metrics

[gweihir]

Nobody defeated humans at Go. There was a rigged stunt (the machine had many, many games the human played to prepare, the human had none and even after only seeing the machine play 3 times thought he could come up with a strategy to beat it), and there is a very good reason we did not see more games. The machine would probably not have stood any chance after a while.

Re:Unsuitable metrics

[K. S. Kyosuke]

There was a rigged stunt (the machine had many, many games the human played to prepare, the human had none and even after only seeing the machine play 3 times thought he could come up with a strategy to beat it), and there is a very good reason we did not see more games.

That is absolutely not what's happened with computer Go. I find it hilarious how you twist easily verifiable facts into lies to justify your denial.

Re:And why, may I ask

[Brett Buck]

The UN goal is to advance globalism.

Re:And why, may I ask

[PolygamousRanchKid]

is the U.N. spending our money figuring stuff like this out?

When we replace the entire U.N. with AI, it won't do stuff like that any more.

Re:And why, may I ask

[K. S. Kyosuke]

United Nations

You don't say!

Re: And why, may I ask

[dougdonovan]

the one with the most toys always wins.

Chance of accusation

[hackingbear]

U.S. President Donald Trump has accused China of stealing American innovations and technology and has slapped trade tariffs on $234 billion of Chinese goods to punish Beijing.

So how is this story really related to the Chinese IP theft accusation? Perhaps, we should also be reminded of American's own dirty history and current activities?

Actual ai

[argStyopa]

Actual AI would be such a colossal game changer, I'm less worried about those who've published the most. Rather, I'd be concerned about those working in secret that achieve a breakthrough.
Its task number one, logically, would be: how do we prevent anyone else succeeding at this?

Re:And why, may I ask

[MightyMartian]

Because nationalism and regionalism lead to such a better world. I've got a few world wars (the Thirty Years War, the Napoleonic Wars, WWI and WWII) to show where that road lead. The whole point of economic entanglement is to make general war between the Great Powers unthinkable. A review of Churchill's post-war speech on a united Europe, not to mention his spearheading of the Anglo-American alliance (via the Atlantic Charter in 1941), should be studied by anyone wanting to understand the precise underpinnings of what we call globalism.

Can it be improved? To be sure. No institution, agreement or alliance should ever be permitted to float on on its own momentum, but neither should it just simply be rejected because of some suboptimal results. These set of agreements created the conditions for a degree of economic growth and general improvement of the people living in those countries never before seen in world history.

Re: And why, may I ask

[olsmeister]

Great. My nephew wins then.

Re:And why, may I ask

[AHuxley]

So the UN can give free money to 3rd and 4th world nations to build their own AI.
The AI will give advice on what loan to accept and what to export from the 3rd/4th world nation.

Re:And why, may I ask

[AHuxley]

The UN given AI will be a nice boondoggle.
All the experts needed to help set up the AI in a 3rd/4th world nation.
The NGO and charity work to support the AI. The education of how to work with and learn from the advanced AI.
The local staff working to help the AI get a nice set of UN SUV. For the big new AI project.
Engineers will be needed to set up the AI building. Local staff will have to decades of paid work with and around the AI.
The CIA and MI6 get to become AI "experts" and live in the 3rd/4th world nation with a new academic cover story.
China will offer a free AI with a road, bridge and port loan.
The CIA and MI6 get to spy on the free AI from China.
Russia offer free academic AI with powerful fast free fully encrypted "academic" network back to Russia.
French experts offer a free Ada super computer with decades of free French educational support in French.
Its not yet an AI. Its 100% designed in France and can do very advance math quickly in all tropical conditions.
French academics will support the computer system over decades for free.

Everyone wins with the sale and support of an AI to a 3rd/4th world nation.
Jobs, spying, engineering, a nice new SUV, NGO work, roads to the AI site.