When I have heard Kai-Fu Lee speaking about this he tends to focus on deep learning in the business-to-consumer setting. It is certainly true that the US and China are currently ahead of Europe in the B2C setting, but I don't believe this to be true in B2B, B2PublicSector, etc. Europe produces far more PhD graduates than the US. Some statistics I saw recently - I can dig out the source information - stated that there were more AI IPOs in Europe than in the US in 2018. If you look at some of the biggest names in AI (Hinton, Thrun, LeCun, Ng, and others) they are European. 

I also don't agree with Kai-Fu Lee's position that European focus on regulation, especially around the protection of personal data, is a liability. I would argue that it is an advantage. Personal privacy, agency, free democracy, etc., are core values in Europe, and are therefore core to what we expect from technology. 

Europe certainly has challenges, such as keeping its talent pool in Europe (or attracting the diaspora back to Europe), a strong funding culture for innovation, etc., but we can and are addressing those.

Finally, if Europe wasn't a strong market, how does one explain the significant increase in presence of Chinese tech/AI companies in Europe?

The future is very bright in Europe for AI, and Europe is already a leader in specific sub-areas of AI, and that will only improve.

Dear Celestino,

first of all thanks for sharing this interesting interview from Kai Fu Lee. I have read his book about the AI Superpowers and must admit that he is right in some areas. Currently Currently the AI activities in Europe are still too slow and fragmented to really cope with the advanced entrepreneurial culture in the US or the huge centrally organized R&D and governmental directed AI cluster activities in China. There is a long way to go to keep up - but there are first positive signs in various European Memberstates. A lot of tiny startups are developing interesting platforms for pilots in various sectors. The flagship study ARTIFICIAL INTELLIGENCE - A EUROPEAN PERSPECTIVE shows first good signs that Europe has acknowledge the urgency to compile the forces to develop an R&D network across Europe to compete with the Mr. Lee’s ‘AI Superpowers’. 

Europe has to built on it’s available engineering and technology strenghts in Manufacturing, Automobile, Chemical and distributed High Tech industries but clustering the AI activities will be a key success factor. 

 Best regards

Maik Neubauer (from one of the German High Tech & IT hotspots, Munich)

I am of the opinion that there is still hope for Europe’s artificial intelligence sector. Activities such as the High-Level Expert Group on Artificial Intelligence show that Europe has decided that it is time to accelerate otherwise we will drop off the AI race. We should all put as much effort as possible on all AI aspects - standardization, ethics, research, innovation, etc. There is still time but if we do not put high level sponsorship/support, HR and budget in place asap then it will be getting more and more difficult to be in the front of the peloton.

Yes US and China are ahead of Europe in terms of # of research engagements and startups working in this space. It's a trap to equal an AI venture to a typical digital software venture. By definition an AI system will at some level automate a specific decision making process that humans currently own. Therefore, AI should be trustworthy if it's going to find its place in society. Without trust, without ethical frameworks around it, an AI system will find it incredibly difficult to penetrate and impact society. You cannot engage in research blindly without incorporating these key aspects into your research / implementation work. Hard as it may be - it is absolutely necessary. From that perspective Europe seems to be investing more so than other regions in the world. 

Thank you for opening a discussion on this. Indeed the US, Canada and China are leading the way towards building AI systems. This is evident from the number of publications in high ranking conferences and journals and strong AI research groups residing in those countries. To my knowledge there are no similar groups with available funding and manpower in any European country. Mainly because people wanting to do PhDs in AI are applying to those strong research groups because of their prospects. Indeed traditionally European Univerisities (mainly in the UK) had strong reputation for training people in AI but those people subsequently left because of the financial incentives given to them by institutes abroad.

However, things are looking brighter by big technology companies opening up research labs in Europe (Google Brain, Facebook FAIR) but this has nothing to do with advancing AI research as companies are mainly interested in the business value that such AI systems bring and will often not explore research direction which are of no significant business value in the near future.

In short, there's still hope, the relatively short distances and small time-zone differences should be taken advantage of such that european institutes can form collaborations towards common research directions. It would also be of tremendous help if the EU could announce funding opportunities for research in this direction. (again excuse my ignorance if such funding opportunities are available through EU)

It is definitely not the case that Canada is leading the way in building AI systems. Not sure how we got lumped in with the US and China. I also don't see the correllation between advances in AI and the number of PhDs or papers published. What we need is original, creative thought. Not regurgitation of the same-old, same-old. Lets not go that route. I agree with Celestino that this EU initiative has the potential to be a real game changer.

Dear All - who have contributed to Celestino's request based on the sifted article/interview - I had a short chat with a sifted reporter about the article and she is interested to create a follow up article based on our discussion here. 

Would it be ok for you to be quoted in an sifted article ? 

Pls. send a short feedback to my personal e-mail address maikneubauer@icloud.com.

Thanks!. 

Best regards Maik 

I think everyone agrees that there are a lot of bright minds in Europe, and we can also agree that we want to promote a kind of AI that respects values like privacy, fairness, explicability, etc.

What is lacking in Europe is an idea of where we want to be in 10, 20 or 30 years from now. Where do we want to be leaders? Here are some ideas: IoT/Industry 4.0, healthcare, smart homes/energy systems. 

Next, how can we get better than US/China in those areas? Perhaps suprisingly, I think the race is still very open, and the main issue is that nobody has really figured out how to share data in an intelligent way. The US may be very strong in collecting data about our web behavior and using that to target ads, but they are not strong in bringing together and exploiting medical data either - mainly because it's all completely confidential.

Now, how can we advance? Here is a concrete proposal: limit data privacy protection to differential privacy instead of complete secrecy. This would allow actors to actually pool data in a meaningful way, and instantly make Europe the place where digital health (and a lot of other things) are happening. 

Good morning Celestino,

last day i watched on italian show in tv.

There was was a great dancer called Roberto Bolle, and strangely there a see an aprroach that I think it's the right one.

Over here there the video if you wish to have a text of the song translated ( by Franco Battiato )

https://www.youtube.com/watch?v=XH8C0i51ezk    

https://lyricstranslate.com/it/LA-CURA-care.html

Have a nice day

Luca Cattaneo

The most worrying aspect of the Kai-fu Lee interview is his understanding of how the IT corps are driving modern capitalism with a monopolistic-winner takes all approach. He is by no means alone in this belief as a resent Forbes article well illustrates.[1]  David Silvers discussed Benedict Evans’, a partner at venture capital firm Andressen Horowitz, essay Tesla, software and disruption[2]  and speculated that, ‘Benedict Evans' question of whether autonomy presents a winner-take-all market: maybe. The industry is so young that it's too early to declare a winner yet, and whether that winner monopolizes the industry is not yet clear.’ The belief that in a few decades globally there will at most be just handful of Automated Vehicle (AV) technology winners is quite common among researchers and executives within this sector of the IT industry.

Lee further observes that ‘[t]he Chinese business models are actually more impregnable than the US ones; the tech giants that the Chinese have come up with are much harder to dislodge because of the winner takes all approach in China.’ He appears quite at ease with a world of IT monopolism and gives a bizarre account of a battle between Microsoft and Google as being ‘how you beat one giant with another giant’ and how he doesn’t ‘really buy this story that monopolists can never be challenged.’  After bizarrely suggesting that Europe should ‘kick out Google, Facebook and Amazon’, but the notion notion that Europe might regulate against monopolism is immediately dismissed because it would not ‘monetise’ as well as the Google and Facebook ‘money-printing machines’. He believes that ‘…the idea that Europe can get ahead by limiting the power of these tech giants is naive. It appears unimplementable both because Europe is so far behind and people won’t give up what they love: the benefits of Google. It is a capitalistic world.’

The monopolistic winner takes all approach is not capitalism. Obviously capitalisms is always in danger of being corrupted and destroyed by the monopolists, and throughout its history there have been numerous attempts, some more successful than others, at limiting the powers of the monopolists For sure, there has been and still is ill-conceived legislation that overly extends and strengthens monopoly power, as in the IPR legislation by the EU. There may well be a problem that people may not want to give up or limit their access to whatever a monopoly is providing, but that, in almost all cases, cannot be used as an argument for monopolism and against strong antitrust laws.[3]      

I do not believe Kai-fu Lee is so naïve to believe that what he is describing is a ‘capitalistic world’. Perhaps he thinks that we are so naïve that we not only believe it is capitalism but that it is the new capitalism that must never be regulated even when it is disintegrating into little more than a plutocracy . It is not new, for it is much the same capitalism that crashed in 2008 that is still being bailed out by the citizens of Europe and elsewhere.  But again, Lee claims, there is nothing Europe, with its admirable ‘idealism’ and protection for privacy, can do about the clash of the IT titans in their capitalistic world but ‘build a brand new application and focus on developing that’ and use ‘idealism to find the next paradigm shift and apply [its] innovation there.’

What should perhaps concern us is that this distorted vision of capitalism is quite common among those that seek to promote IT so-called ‘AI’[4]. It should perhaps be noted that Lee’s  vision of the IT industry’s money machine business model is already dated and is showing signs of a correction that hopefully will become major and long-term. However, for reasons that go beyond this brief discussion piece, much of the response from governments, business, academics, media, public and of course the EU to this vision has been for the most part an uncritical acceptance of the hype and predictions by researchers and the tech giants that ‘AI’ will be massively “disruptive”.[5] As with the first ‘AI hype’, there is no credible independent evidence that the (narrow) AI presently being developed will have any greater impact on a global scale than any of the other peaks in technological change since the 18th century. For example, deep learning convolutional neural network (CNN) should greatly increase the quantity, speed and accuracy of diagnosing many conditions and will improve healthcare.[6] However, we would have to be remarkable ignorant of medical history not to realise that technology has been effectively doing this for centuries and that CNN and IT systems should only be seen as another major advancement in medical technology. AV technology is at the research stage but cannot possibly safely and securely deliver a fully functional and economically viable AV system at SAE level 5 in the foreseeable future or indeed, as John Krafcik CEO of Waymo says, ‘ever’[7]. AVs could eventually be effective in some urban areas and on motorways, but the notion that they could or indeed should be the major go anywhere anytime road transport technology is not only naïve, it is dangerously naïve.

So in brief I believe that today’s ‘AI hype’ is being fuelled by a number of factors. A misunderstanding of technology - not just IT but all technology - is one. In the case of technologies like AVs, of which there are many, they are often presented as the answer to a problem that upon closer examination is not a problem, or if it is a problem ‘AI’ is not the answer[8]. ‘AI’ is also not only claimed to be ‘disruptive’ but unprecedentedly disruptive and indeed could bring humanity to an end[9]. It is all pretty much the same as the first ‘AI hype’ with the exception of the monopolist winner takes all descriptions of capitalism. Fifty or so years ago it was for the most part the military–industrial complex that saw ‘AI’ as a winner takes all technology. Now, as Silver, Evans, Lee and others would have us believe, we are confronted with the spectacle of tech giants slugging it out in their own virtual reality game they have dubbed “Capitalistic  World.” I tend to agree with Adam Smith, that without proper regulation of capitalism and state intervention to protect people from the detrimental effects of the division of labour (aka ‘machine intelligence’), it will eventually stupefy and degenerate humanity. His ‘eventually’ may be our ‘now’, so we might be too late. 

(Sent from phone) 

[1] https://www.forbes.com/sites/davidsilver/2018/09/04/self-driving-cars-w…

[2]https:/www.ben-evans.com/benedictevans/2018/8/29/tesla-software-and-disruption

[3] As Smith explained in the Wealth of Nations, the state can and should operate a few monopolies (his main example was the post office and some transport) because they are more efficient than the free market.  

[4] I have rejected the term ‘AI’ in my research since the 1980s because it is hopelessly misleading and was, as John McCathy admits, a term he invented to get funding. Narrow or weak AI is only a type computing, information processing or information processing (all of which fall under McCathy’s definition of AI), so I will use the acronym IT where others might confusingly use AI. Where I do use ‘AI’ or ‘AI hype’, it will be for the purpose of emphasis.

[5] Up until about 2009, AI was not described as being “disruptive” by followers of Bower and Christensen’s theory of disruptive technologies.  Bower, J. L., and C. M. Christensen. ‘ Harvard Business Review 73, no. 1 (January–February 1995): 43–53.

[6] CNN machines can never operate perfectly and will make errors, so, as with all machines, there will always be an important role for humans in the “system”.

[7] Even at SAE level 4 the system would regularly “crash” because of weather and road conditions, and all AV levels would regularly crash for much the same reasons all IT systems crash. https://www.cnet.com/news/alphabet-google-waymo-ceo-john-krafcik-autono…

[8] The one of the major problems is claimed to be that 90% of road fatalities are caused by human error.  Of the remaining 10%, a percentage will be caused by soft and hardware failures, which will of course greatly increase with the number of AVs on the road, all of which will be far more complex, subject to security attacks and prone to considerably more soft and hardware failures, many of which will cause large scale systems failure, than today’s automobiles.

Another problem is that road fatalities in US, adjusted for deaths per 100k. km., road type, etc., are two or three times more than most of European. Continued technological development, road improvement, drink drive preventive measures, etc., would reduce European fatalities below that of AVs system.

The problem of road haulage cannot be solved by AV technology because of the need for drivers

during loading and unloading.

In short we should critically ask whether AV technology is answering any major problems and not engaging in

idle speculation about ethical problems.

[9] IT research in the area of so-called Artificial General Intelligence (AGI) is going pretty much nowhere. Indeed, the Artificial Intelligence Index: 2018 annual report shows no real advances in this field. It is still reporting closed domain game playing as ‘AGI’ advances, which is a sign of desperation.   

http://cdn.aiindex.org/2018/AI%20Index%202018%20Annual%20Report.pdf