Can AI beat all stock markets

Robotic fund : How investors can benefit from the use of artificial intelligence

Many think of the film "Matrix" or of humanoid super robots in a distant future. In truth, however, Artificial Intelligence (AI) is already present in many places in our everyday life: Speech recognition systems such as Alexa from Amazon or Siri from Apple work with machine learning, as do translation programs, software for fraud detection and data analysis for security monitoring.

This trend has also arrived in financial investments: On the one hand, many fund companies are now offering securities that investors can use to invest in a basket of companies that deal with AI or benefit from it. On the other hand, many fund providers now use a machine, at least for parts of their knowledge and selection processes. For some providers, the digital fund manager is now even a reality.

Tagesspiegel Background Digitization & AI

Digital politics, regulation, artificial intelligence: The briefing on digitization & AI. For decision-makers and experts from business, politics, associations, science and NGOs.

Free test now!

However: the term “artificial intelligence” is misleading. So far, it has consistently only been about so-called weak AI, in which computers sift through huge amounts of data, calculate options based on previous developments and thus relieve people of work that they could only do incompletely and in much longer periods of time. Highly specialized software is trained by processing masses of data for a small work area in order to later recognize similar patterns and draw conclusions from them.

AI experts expect a benefit here, for example, in medical diagnoses, in which computers can screen data or tissue much more precisely, in speech recognition, in the networking of machines with one another.

So far only "weak AI" is in use

In 2017, the DeepMind software, which belongs to the Google parent company Alphabet, caused a sensation and with an artificial neural network - equipped only with the rules of the game and a reward system - learned the rules of the highly complex game of Go within a very short time and is now unbeatable. The system did not learn from previous human games, but independently and continuously calculated those with the highest probability of winning from billions of moves.

The fact that the developers gave reward points for advantageous moves accelerated the process. However, the following still applies: There is currently no real AI that is equal to or superior to humans and that would learn independently and in parallel in many fields and in all areas of life.

Fund companies also use weak AI or simply software that is specifically programmed for certain results and can only learn independently in this area. So far, the success has been modest. Funds whose shares are equipped with the help of the weak AI mostly do no better than funds whose shares have been selected by a person, or even no better than the overall market.

The funds of the fintech Catana are not performing

At Catana, a German fintech for “technology-based asset management”, computers sift through the opinions, sentiments and historical price movements of around 45,000 securities worldwide in real time. They evaluate two million messages and texts every day - and thus come up with recommendations for action. The entire process from viewing the data to executing the order is automated; German, English and Chinese are processed, according to Catana.

In the end, only one person, the fund manager, confirms with a click what the computer has calculated. The "Data Intelligence Fund", which has been operating according to the computer specifications since November 2018 and is not only allowed to invest in shares from Europe, but also in financial constructs such as swaps and futures, has so far not been able to convince. The performance has been around zero since March. The fund is also expensive: In addition to a front-end load of up to six percent, there are regular costs of 2.12 percent and profit sharing. And: In 2017, Catana had already withdrawn a similar fund after heavy losses.

AI sales could increase tenfold

Allianz Global Investors also devotes itself intensively to the topic of AI. The fund invests primarily in US companies (91 percent of the portfolio), as well as in companies from the Netherlands, France, China, Japan and Korea. Top positions are the streaming pioneer Roku, the semiconductor and storage companies Broadcom, On Semiconductor and Marvell, the payment service Square, Microsoft, the cloud communication company Twilio and the online credit facility Lendingtree - all companies that have so far only used artificial intelligence Fringe plays a role.

However, with an increasing proportion. A study by the management consultancy McKinsey predicts that by 2030, 70 percent of all companies would use at least one technology from the field of artificial intelligence. By then, this could create an added value of $ 13 trillion and increase global economic output by 1.2 percentage points per year.

In the opinion of Hans-Jörg Naumer, head of capital market analysis at AllianzGI, AI could change the world far more than the Internet. The growth forecasts for AI applications are enormous: Experts such as Naumer believe that sales with AI could increase almost tenfold from currently 3.7 billion dollars by 2025. Companies that jump on the bandwagon first could benefit. So far, the very young funds have not yet reflected this megahype forecast.

Allianz GI has also started to involve machines in sifting through data and selecting stocks. The software helps to sift through large amounts of data, recognize structures and prevent errors, explains company spokesman Klaus Papenbrock. In fund management in particular, however, it will remain a coexistence of man and machine: "Artificial intelligence, that is clear, is not able to look into the future."

The Oddo AI fund aims to beat the MSCI World

In fact, the stock market lives primarily from uncertainties, from the fact that future events and developments cannot be reliably foreseen, but are often dominated by chance, or that market participants judge facts differently - a barrier that neither human nor artificial intelligence can cross is. When it comes to stock selection, AI users hope that the machine-based screening of huge amounts of data on companies, sentiments and opinions will lead to information advantages that the competition does not have.

However, it is uncertain whether this could also lead to a price increase, i.e. added value. Because stocks are not only controlled by measurable facts, but above all by spontaneously occurring emotions such as greed, fear or hope.

The German-French asset manager Oddo BHF nevertheless hopes to use artificial intelligence to beat the MSCI World index in the future, at least within a period of five years. The stocks for the Oddo Artificial Intelligence fund are selected with the help of a machine, the computer sifts through millions of data every day and searches for relevant moods and trends. In the end, 60 shares have been determined, all of which deal with the topic of AI. The fund manager reviews the process on an ongoing basis. In the event of success, the investor will also be asked to pay.

To home page