By Hagen Ernst, Deputy Head of Research & Portfolio Management
AI, take over please
Artificial intelligence (AI) is considered to be a key technology of the future - and not only since the digitalization push caused by Covid-19. Its fields of application are endless - and its growth potential enormous: market research companies like Tractica expect a tenfold increase in sales by 2025. Thanks to faster computing power and better data quality through the Internet of Things and 5G, the foundation of the AI era has been laid. An investment theme for the DJE - Alpha Global.
We are still far from machines that can beat the human intelligence. But: AI applications can already solve specific problems today based on available data. Unlike rule-based systems programmers do not instruct AI-based processes how to react to user steps. Instead they define algorithms that independently learn to find patterns and make decisions. An AI system has neither intuition nor brain and therefore has to be trained by the use of many examples - i.e. enormous amounts of data. This is also called machine learning. AI is superior to rule-based systems in that it can react to new situations and learn from experience.
Machine learning and Edge AI on the rise
Deep learning is a sophisticated method of machine learning. This means that AI systems use artificial neural networks to imitate the incredibly complex brain, which consists of over 80 billion interconnected nerve cells. During the learning phase, an artificial network that is supposed to recognize faces, for example, is fed a vast number of facial images and confirms a correctly recognized face each time. In the case of an "OK", the network reinforces the connections that led to the correct result. Those that led to the wrong result become weaker. After many attempts, the system becomes more intelligent and continues to improve on its own. On this basis, the Watson AI system from IBM has learned to recognize cancer cells. Artificial intelligence also enables the recognition of feelings and speech, automated driving or translations in seconds. Until now, AI has primarily taken place in powerful supercomputers or in the cloud. But the so-called Edge AI is increasingly gaining in importance. Here, the data is processed decentral, i.e. directly in a terminal device. The advantage: there is no transmission time.
From cloud business to autonomous driving: US tech giants ahead
The market for high-performance processors is shared by the two U.S. chip developers AMD and NVIDIA, with NVIDIA leading the way in market share (currently around 70 percent) and performance. With the AI chip TPU2 the US holding Alphabet also owns a high-performance processor that was developed together with the US semiconductor company Broadcom and is strongly involved in the further development of AI. In addition Alphabet's subsidiary Waymo is a leader in the development of autonomous driving, alongside Tesla and the Chinese Internet company Baidu. And: Alphabet is particularly advanced in AI-based speech recognition and the use of AI in the cloud business. The material, i.e. the masses of required data, are supplied by social media, search engines and online commerce, among others.
Chinese tech giants are advancing into the FinTech sector
Pure online payment systems were only the beginning. The online payment service PayPal uses AI to uncover fraud attempts in the payment system and other banks and financial service providers evaluate credit default risks. Chinese tech giant Tencent, for example, wants to grow primarily in its financial services sector and, along with Alibaba, is already a leader in payment processing. Via WeChat, the Chinese WhatsApp, the group collects extensive data and uses it to evaluate payment behavior and credit risks, among other things. Access to customer data and big-data expertise should give Tencent a competitive advantage over traditional providers. In the medium term the company aims to become the largest provider of customized financial services in China.
AI redefines asset management
In the future, AI could independently control an entire asset management. BlackRock, the world's largest asset manager, already operates the Aladdin analysis system via four data centers. This system processes a daily growing amount of historical data and selects the most probable scenarios from the multitude of possible scenarios. The result is a statistical picture of future scenarios for securities portfolios under various future conditions. Stress tests can be performed, for example, by simulating how a bankruptcy à la Lehman Brothers in 2008 would affect a portfolio. At present, however, Aladdin is still a long way from being a purely AI-based asset manager and still needs a portfolio manager of flesh and blood. But the advantages of an AI system are obvious: It can process vast amounts of information very quickly - and emotions have no influence on its investment decisions.
Healthier thanks to AI
AI is also likely to make enormous progress in the healthcare and pharmaceutical industries. The development of new drugs is becoming more and more complex: development costs double about every 5 years and 90 percent of all new developments are not even approved. Here, an appropriate AI system could make a pre-selection in the shortest possible time, for example by analyzing millions of molecules for their specific suitability as a drug against a certain disease. NVIDIA, for example, plans to put the UK's new Cambridge-1 supercomputer with a computing power of 400 petaflops into operation by the end of 2020. A petaflop is 1,000 trillion "floating-point operations" or mathematical operations per second, which enables about three million times more computing operations than everyday computers. The pharmaceutical companies GlaxoSmithKline and AstraZeneca, which are both working on the development of a c Corona vaccine, will be the first to have access to the Cambridge-1.
AI systems such as IBM's Watson are already in use for medical diagnostics. It recognizes and classifies cancer cells, for example. Watson will also optimize established types of diagnostics and accelerate basic medical research such as genetics and cell development. AI is also rapidly advancing genome research and has, in particular, catapulted forward Next-Generation Sequencing, a high-throughput method for decoding DNA. Less than 20 years ago the complete sequencing of the first human genome swallowed up three billion euros and took 13 years. Today, genomes are decoded within days.
Autonomous driving - just a question of time
AI will make autonomous driving possible. A large number of cameras as well as lidar and radar sensors collect data from the environment and a high-performance processor in the car will process them in real time. This means that autonomous driving on the highway or in traffic jams is already possible today, as is automatic parking. Although we are probably still decades away from real autonomous driving, the foundations have been laid. Traffic flows could be significantly optimized by the interaction of networked vehicles, traffic observation systems and the communication of navigation systems with each other. A large number of companies from the old and new economy are currently working together to achieve this: Car manufacturers such as Daimler, BMW and Tesla, suppliers such as Continental and Bosch, semiconductor manufacturers such as NVIDIA and Intel, and technology giants such as Alphabet and Apple. Alphabet with Waymo, Baidu and Tesla are probably the closest to the goal of an autonomous driving car. Among the high-performance processors, NVIDIA and Intel systems seem to be gaining ground.
Transformation of the Old Economy
Thanks to Big Data, AI can be used in more and more areas of life: From speech recognition, robotics and marketing to banking, customer care and mobility, telemedicine and cyber security - many old economy industries will have to change sooner or later. For example, pilot projects of the first AI-based factories at BMW or Kion, a German warehouse technology provider, show that the fully automated, flexible operation of logistics centers works. Robots take over the production and the various processes, and humans merely train the robots and monitor the processes.
Two industries are feeling the effects of the change particularly keenly: the automotive industry could lose its status if more efficient and cost-effective mobility concepts are created with the help of AI, and if the own car becomes less attractive. In contrast, Alphabet, with Waymo and the map material from Google Maps, could become a competitor in autonomous driving. Traditional financial service providers could also come under pressure: FinTech service providers are no longer limited to payment services; they also offer solutions for investments, consulting and credit services.
Transformation of the Old Economy
Conclusion: AI is anchored in everyday life - and a multifunctional growth driver
AI is already used in many areas such as production, quality control, logistics, customer care or telemedicine. In everyday life Alexa or Siri, for example, support us as personal digital assistants - and AI systems make it possible to park and vacuum without any action on our side.
In the future AI is likely to find its way into more and more areas of life and work and will play an increasingly important role across all industries. An exciting and complex investment theme for the coming years.
Note: All information published is for your information only and does not constitute investment advice or other recommendation. Long-term experience and awards do not guarantee investment success. Securities are subject to market-related price fluctuations which may not be compensated for by the active management of the asset manager or investment advisor. This information cannot replace a personal consultation. All information has been provided with care and to the best of our knowledge at the time of preparation. Despite all due care the data may have changed in the meantime. Further information on opportunities and risks can be found on the website www.dje.de. The sales prospectus and further information are available free of charge in German from DJE Investment S.A. or at www.dje.de The fund management company is DJE Investment S.A. DJE Kapital AG is the distribution agent.