The Artificial Intelligence Industry and Global Challenges

Whoever controls the strongest artificial intelligences controls the world.

Artificial intelligence is the most important technology of the 21st century. It is therefore important to understand global ambitions and movements.

In this article I examine the global artificial intelligence industry and in this context consider the aspects of politics, data, economy, start-ups, financing, research and infrastructure.

I will only briefly discuss the current superpowers China and the USA, as I will dedicate a separate article to each of them.

The question that we must ask ourselves in the end is how humanity will deal with the global challenges.

Artificial intelligence needs more attention from politicians

So far, the first wave of digitization has developed without much government influence. Although there are now plans to break Google’s monopoly (USA and Europe), for example by imposing European fines on Google and Facebook, politics is lagging behind the market by over a decade.

As far as AI is concerned, for the first time in recent history I have observed a multitude of initiatives, strategies and actions by dozens of governments around the world — with very different goals and approaches.

Artificial intelligence is and remains an issue that politicians and administrations of all nations have to deal with.

AIs are relevant for climate protection and economic policy.

AIs influence the governance of domestic industry, the security and privacy of citizens.

A long-term strategy for the establishment and development of own AIs is crucial. But it is also expensive. Europe in particular has problems deciding in favour of long-term and investment-intensive strategies.

China has a clear vision of how country wants to master artificial intelligence. From China’s point of view, artificial intelligence is an important tool for strong foreign policy, military dominance, economic success and for controlling one’s own population.

The USA benefits from a strong research cluster and the super corporations Google, Microsoft, Facebook and Amazon, each of which is in the lead of the AI development.

Although the USA has not yet found a red line under President Trump, the state has been promoting the research and implementation of AIs for decades through its countless secret services and ministries.

Canada and Israel have become equally important but smaller players in the global competition for AI rule.

Israel, always very technologically strong, has more AI companies than Germany and France put together (see also our study Global Artificial Intelligence Landscape). In Israel, there is a close network of universities, access to the Asian and American capital markets, close cooperation with the military and the government. The Israeli company Mobileye was bought by Intel for $15 billion and is just one example of a thriving AI ecosystem.

Canada benefits greatly from the renaissance of deep learning in the last 7 years. Geoffrey Hinton, Yann LeCun and Yoshua Bengio are three of the strongest researchers in this technology. All three have researched at different times in the Canadian Institute for Advanced Research. Together they have survived the last “AI winter” and have been shaping the market ever since.

In addition, Canada has a clear AI strategy, research, investment and implementation have been promoted for years.

Also worth mentioning are Japan, Korea and India, which have good prerequisites for playing a relevant role in the AI industry in the coming years.

A reading reference at this point is the report of national strategies of artificial intelligence of the Konrad Adenauer Foundation (Part 1 and Part 2).

Economic power through artificial intelligence

While politics provides the framework conditions for research, financing, education, data, promotion and regulation, in the medium term AIs must be developed by companies and brought onto the market.

First of all, national interests have to be taken into account.

These include, often with their own agenda and independently, global corporations with their own AI research and AI products.

In my view, Google (Alphabet), Amazon and Microsoft are global leaders. The Chinese Internet giants Alibaba, Baidu and Tencent are also relevant players.

There are two types of companies: Those that develop and sell AI as a core product and those that use AI to complement their value chain.

Either way, any company active today has to deal with artificial intelligence. On the one hand, AIs can replace existing business models, and on the other hand, they can be integrated into countless company-internal processes: Accounting, controlling, production, marketing, sales, administration, personnel management and recruiting.

By the way, this is the primary driver of applied artificial intelligence: reduce costs and maximize profits.

And, of course, it’s also about control. Every AI used takes over activities that were previously performed by humans. Often, after a while of training, the AI is faster, more efficient and cheaper than the human being was before.

People become ill, they need holidays, food and sleep. They have to be entertained, quit or retire. AIs work 24/7 and do not demand a wage increase.

The more companies use AIs, the more independent they become of human labour.

Data is a competitive advantage

The foundation of any artificial intelligence is data. We therefore need data on several points.

First of all, we need data for the research and training of narrow artificial intelligences. The more digital your business model is, the more data you have.

For this reason, marketing leaders (Google, Facebook), software companies (Salesforce, Microsoft) and e-commerce retailers (Zalando, Amazon) have been heavily involved in AI for years.

Some banks also recognized the trend early on. Therefore Goldman Sachs and J.P. Morgan have already recruited thousands of employees with a focus on machine learning and data science.

Those who have their own data can achieve an enormous competitive advantage.

Those who have no data have to collect, store and evaluate data.

However, this is where the different national data protection laws come in, which is why Europe is at a disadvantage.

GDPR/DSVGO may indeed have the good intention to create a European data internal market, but currently form an enormous location disadvantage for Europe.

The fear of the regulation paralyzes whole industries. Personal discussions with clinics and doctors showed me that the health industry no longer shares any data. This literally costs human lives, because this obstacle is detrimental to health research and life-prolonging algorithms.

This is just one example among many.

Uncertainty about data is paralysing our entire European industry. For fear of penalties, data is not collected at all. We are creating a culture of data anxiety at a time when data is actually our strength.

Europe is the most important data market in the world, but we are wasting our potential.

China, on the other hand, is the extreme opposite. The state helps with a lively exchange and centralization of data (more on this in the chapter on China). In addition, the population has fewer concerns about the free handling of data.

De facto, privacy no longer exists in the 21st century. Every digital action is measured and stored. However, we Europeans are sticking to an old ideal.

Artificial intelligence start-ups are the giants of the day after tomorrow

Start-ups are essential for any economy because they take on two essential functions of an ecosystem.

Start-ups are drivers of innovation. These young companies are often more courageous, faster and more flexible in developing new products than established companies. Backed by the capital of venture capital funds and business angels, start-ups take high risks in the expectation of extraordinary success.

Although 95% of start-ups do not survive the first 5 years, the entire ecosystem benefits from them.

Companies can buy new products and innovations through acquisitions.

Former employees find new jobs and transfer their knowledge.

Investors and founders learn and take their knowledge with them into new projects.

Perhaps the young company will survive the 5-year threshold. It secures financing (from seed to IPO), gains talent, grows, develops products for which customers pay, scales and becomes a corporation. Facebook, Google, Apple, Amazon, Uber — all started out as start-ups and are now dominant market leaders.

Charles-édouard Bouée, former CEO of Roland Berger, said at the 2018 Rise of AI conference that the next wave of trillion-dollar companies will mainly be AI companies.

This won’t work without start-ups. That’s why we need to encourage building start-ups.

Research is more important than ever

The rediscovery of Deep Learning was only the beginning. The field has evolved through new approaches from CNN, GAN to evolutionary algorithms (Prof. Damian Borth’s presentation at the 2017 Rise of AI conference is a good introduction to deep learning).

Computational linguistics around NLP and NLG has also made enormous leaps.

Today, hundreds of thousands of narrow artificial intelligence applications are based on the research results of the last 30 years, after we reached the critical volume of computing power and data availability in 2012.

Where do the research results come from?

On the one hand, they come from universities. MIT, Stanford, Carnegie Mellon University and Berkley are lighthouses in AI research (see also the AI index from Stanford).

MIT alone is investing 1 billion dollars in the training of new AI degree programmes by 2020.

On the other hand, companies have now become a major driver of AI research. You should know Google DeepMind. Microsoft has over 8,000 AI researchers.

Leading minds conduct research for corporations with more data and financial resources: Richard Socher (Salesforce), Yann LeCun (Facebook), Andrew Ng (until 2017 Baidu) or Demis Hassabis (Google).

European universities and corporations, on the other hand, are not leaders in the field of AI research. Of course, we also have smart minds like Prof. Jürgen Schmidhuber, Prof. Francesca Rossi and Prof. Hans Uszkoreit.

In addition, there are AI courses at KIT, TU Munich, TU Berlin, the University of Osnabrück (Cognitive Science), Oxford and Cambridge University.

But all this is just mediocrity and not internationally recognized top-level research.

Instead, the DFKI (German Research Institute for Artificial Intelligence), dozens of Max Planck Institutes and Fraunhofer Institutes in Germany in particular are primarily engaged in applied research. But even these institutes do not manage to play in the first league in the global competition for talent, data and capital.

But it is precisely research that will be decisive in the coming decades when it comes to the question of who will develop the first general artificial intelligences.

Video recommendation: Lecture by Prof. Hans Uszkoreit at the Rise-of-AI Conference 2017 on Super Intelligence.

Without infrastructure there is no artificial intelligence

By infrastructure I mean not only the availability of data but also the necessary computing and performance capacities.

NVIDIA used to be known for their graphics cards among gamers. Today, NVIDIA is one of the leading manufacturers of GPUs, which are increasingly used for AI applications. Google, Intel and many other companies are very active in the development of new AI chips in various forms.

At the same time, Microsoft, AWS, Google and IBM are expanding cloud capacity around the world to meet growing demand.

While China will focus strongly on 5G, which is critical for real-time AI applications and the networked industry, Europe will not play a leading role in this technology issue either.

Artificial intelligence must be financed

The development of artificial intelligence is expensive.

Top AI researchers are rare and receive salaries of up to €300,000 per year.

Data must be collected, sorted and labelled. Developing AI models takes time for experiments, mistakes and new methods.

AIs need data, must be trained and educated.

These costs are borne by companies, start-ups, investors and also the state.

China has understood this and is investing over 130 billion euros in the Chinese AI market. Provinces such as Beijing, Shanghai and Tianjing are each investing tens of billions in local AI industry.

In the USA, Google, IBM, Microsoft, Amazon, Facebook and Apple have already invested over 55 billion dollars internally by 2015.

Without money, there is no artificial intelligence.

And once again Europe is too stingy to invest in the future.

A comparison of the orders of magnitude: In 2018, the German Bundestag had budgeted as much as € 500,000 for AI funding. A further 500 million is planned, but the funds are not yet available.

Progress will not succeed in this way.

At the same time, China is financing 400 new chairs for AI. To date, we have seen nothing of the 100 new professorships planned under the German AI Strategy.

In this context, I would like to praise Great Britain because it is going against the trend in Europe — despite Brexit. More money is being made available on the island for start-ups and universities in the field of artificial intelligence.

If you want to know more about the current state of AI, I recommend the State-of-AI-Report 2019 and my presentation of the Rise of AI 2019 as video.

Where does Europe stand in the global AI arms race?

As I mentioned earlier, Europe is currently losing the competition for the leading AI nations.

While Europe is still considering whether to compete at all, China, the US, Israel, the UK and Canada are already competing for data, markets and talent.

Our problems in Europe are homemade, they are the result of our inertia, lack of vision and ambitions.

There is a lack of money for education. Not only are our schools and universities underfunded, but so is the education labour market. Our children are not learning enough about digital skills. Our students rarely take AI-relevant subjects. Our working population lacks retraining opportunities that also meet the needs of the growing digital industry.

The transfer of research results to industry is sluggish. Results either disappear into the drawer, or the IP transfer is in bureaucratic terms a horror, especially for young companies and spin-offs.

Our European AI start-ups are significantly underfinanced. Those who currently need money from investors must market e-bikes and e-scooters, but they should not include technology. The more complex the product, the more difficult it is to get capital. The simpler the business model, the faster the accounts are filled.

Although many talents from Asia and America want to work in Europe, it has become bureaucratically complicated. Since the wave of refugees, the offices have been overwhelmed. It is almost impossible to hire talented AI developers from Iran, Russia or China. There is currently a spirit of rejection rather than openness in Europe.

Europe lacks a single strategy. Countries such as Finland, Sweden, the Netherlands or France have their own AI strategies and, moreover, a great deal of ambition. Germany, in particular, is blocking a common European approach and thus possible success.

When I was with the European Commission in 2018, a Bulgarian researcher said that she would be happy if her country had a plan at all. According to her, entire sections of Europe are significantly worse off than we are in Western Europe.

I am not saying that politics must solve all our problems. Companies still have to build products, founders have to start start-ups, VCs have to finance these start-ups and researchers have to do research.

But politicians can support us with a clear strategy. It can build up regulatory structures instead of inhibiting them. It can create incentives for investment and act as a role model. And it must be a matter of course for politicians to take care of the education of pupils, students and qualified further education in general.

On paper you can read all this (AI strategy of the German Federal Government), but in practice nothing happens.

Europe is marked by power struggles, egoism and technology phobia.

Global challenges for humanity

But Europe is only part of the world and must adapt to a global power order.

Therefore, there are many challenges for the growing artificial intelligence industry.

First of all, there is the issue of data protection. Which standard will prevail? Europe has currently set clear standards that will ensure that companies develop their AIs outside the EU.

Will we completely abandon data protection in the future and have Chinese free movement of data? Or will the European system prevail in the medium term?

Artificial intelligence needs supervision

Furthermore, all governments must think about the regulation and supervision of artificial intelligence. AIs are increasingly influencing the media, industry, education, security, military and financial markets. There is therefore a need to regulate how artificial intelligence (and the companies that design it) is controlled.

For example, the Chinese company Squirrel AI helps millions of students to develop individual learning content that matches their learning pace and skills. But who in Europe would control and supervise these AIs in terms of content and subject matter? I cannot imagine that thousands of local education ministries are currently in a position to do so.

Authorities must therefore recruit the necessary specialist staff, develop concepts and put them into practice. This takes time and should happen sooner rather than later.

It needs an artificial intelligence ethics framework

Artificial intelligences make decisions every second and almost every decision has an ethical component. AI ethics and AI morality are inseparable from the research and application of artificial intelligence.

Artificial intelligence can prevent — or reinforce — prejudice, racism, corruption and sexism.

Therefore, an AI ethics framework is urgently needed. At the upper level, every cultural area (often organized in states) has to deal with this. What kind of AIs do we want? Which values should they represent?

This social discourse must be actively started and conducted. Currently, AIs are being developed without controlled morality. It is up to the developers to decide how the machines act later. But society itself should have this sovereignty over right and wrong.

Therefore, every nation, every state system and every ethnic group must begin to discuss their own AI ethics frameworks.

The same applies to companies. Every company needs an AI ethicist, just as there are data protection and equality officers. The AI ethicist ensures that the data is unbiased and that the AIs do not discriminate against anyone.

This also requires AI ethics frameworks in companies that are based on social values and also reflect the corporate culture in the code.

The rise of strong artificial intelligences

We also need to address the issue of growing intelligence and the influence of artificial intelligence.

OpenAI recently received another billion dollars from Microsoft for research into General Artificial Intelligence (AGI), in addition to the billion dollars from Elon Musk.

There is a reason why Musk, Zuckerberg, Hawking and Gates warn against artificial intelligence. Siri may still be stupid today, but could surpass your intelligence in 10 years.

Artificial intelligences become smarter, more knowledgeable, more capable and faster every day. Contrary to our biological nature, there are no limits to the AIs.

Therefore we have to think about AGI, Strong AI, Super AI already today. Also topics such as Neural Interfaces and the Human Operating System have to be discussed.

Enclosed is a video recommendation by Dr. Linde about strong artificial intelligence on the Rise of AI 2019.

The machines are energy hungry

Another global challenge is the energy requirements of the machines. While our brain needs as much electricity as a light bulb, AI applications are extremely energy-intensive.

So if we want to maintain technological progress, we need to solve our energy problem. Otherwise, not all people will be able to live long and healthy lives.

We therefore need energy sources that are sustainable and scalable.

A society without work?

Furthermore, as a society, we must start talking about our work. We Germans in particular are afraid that the machines will take away our jobs.

Personally, I think it would be great if machines could relieve me of work.

But the machines won’t steal your job. I estimate that 50% of today’s human activities are carried out by machines in 20 years because they are cheaper and faster.

This development is therefore a good thing, at the same time it is a challenge.

There will be many new activities that we do not even suspect today. These will include AI kindergarten teachers, AI trainers, AI ethicists and AI controllers.

It is therefore important that we think about how we can retrain those people who perform tasks today that will be performed by machines tomorrow.

It will be the greatest retraining in human history. Some people will not make this change, what happens to them?

Others will use freedom and redesign their lives.

I hope that in the future the meaning of life will not be work, but joy of life. I hope that we get to the point where everyone works because they want to, not because they have to. Let the machines do the work that nobody wants to do. And let us finally pay better the people who take over important social activities such as bringing up our children, teaching pupils and caring for us in old age.

Either way, we need a radical new social model, because the earth does not need 10 billion philosophers, artists, entrepreneurs or programmers.

How do we distribute wealth?

Probably not the last, but a crucial challenge is the question of how to distribute wealth.

If fewer and fewer people are needed for the same productivity, companies’ profits will increase. But companies are often in the hands of a few selected families and funds. Today, 40% of all publicly listed companies in the US are already owned by four gigantic funds.

This trend will result in very rich people becoming much richer. A few chunks of this wealth will then fall to the administrative class (lawyers, bankers, entrepreneurs, investors) and hardly anything remains for the remaining 99% of the population.

Recommended reading: “Capital in the 21st century” by Thomas Piketty.

Civil war or a carefree life?

With artificial intelligence, this trend will become more extreme. I fear that in 30 years around 100 people will control 90% of the world. We are not just talking about money here, but about access to the machine code and thus to the control of the global economy, the military and information.

And what if the majority of people lose their current importance for the system, namely to work and consume? Does the idea of the Club of Rome then become real and we reduce ourselves (involuntarily) to 500 million people?

AI makes it possible.

So if we want to prevent a civil war, prosperity must be distributed beforehand. In any case, so much of it that no one has to fear hunger, homelessness and poverty anymore.

There are many approaches and ideas. Let us please discuss them further, such as the basic income and liquid democracy.

Let us ensure that we live in a world where the machines serve all of humanity.

==

If you like to receive early releases of next chapters, then sign up HERE.

The first 100 people who later order a printed book, will receive a personal copy.