AI vs Humans: Who makes better money?

Artificial intelligence (AI) has now matched or even surpassed humans in areas previously considered unattainable, but can AI really outperform humans at making early predictions? private?

AI Financial Forecasting

The financial technology (fintech) industry is growing strongly with a rapid growth rate. Fintech startups are increasingly challenging traditional financial institutions in areas such as retail banking, grants or personal investments. In particular, AI often plays a role in behind-the-scenes such as cybersecurity, anti-money laundering, customer checks or chatbots.

AI vs Humans: Who makes better money?

There are many areas that use AI and bring success, but one case seems to be absent, such as money in financial markets, it is rarely used in the investment decision-making process. A new study published in the International Journal of Data Science and Analytics tackles the question of whether AI is better than humans at making money.

Quantitative Hedge Funds claim that they use AI in their investment decisions. However, they do not publish official information about the results. On the other hand, academic studies have repeatedly produced highly accurate financial forecasts based on machine learning algorithms. In theory, these could lead to successful investment strategies for the financial industry. In practice, however, that doesn’t seem to be the case.

What is the reason for this difference?

27 peer-reviewed academic studies published between 2000 and 2018 were analyzed. These studies describe stock market forecasting experiments using machine learning algorithms.

Most of the tests run multiple instances in parallel with the investment model. In most cases, the authors presented the most successful model as the main product of the experiment, meaning that the best outcome was chosen and all suboptimal results were ignored. . In fact, this approach will not work in investment management, any strategy can only be implemented once and its result must include clear profit and loss, no has undo result.

Testing multiple variations, and then presenting the most successful one as representative would be misleading in the financial sector and could even be illegal. Only one version of the algorithm should be tested, which should represent a real-world investment setup.

The models in the tests achieve a very high level of accuracy, around 95%, but 5% errors can also be a big problem, wiping out not only profits but also the entire underlying capital.

Most AI algorithms seem to be “black boxes”, with no transparency on how they work, thus making investors less interested. Furthermore, most tests do not take into account transaction costs. While costs have been reduced over the years by no means non-existent, even it can still make the difference between profit and loss.

None of the tests analyzed gave any consideration to applicable financial regulations or issues related to business ethics. The tests themselves do not engage in any unethical practices, so they lack a design feature to ensure that they are in line with ethical standards.

According to analysts, machine learning and AI algorithms in investment decision-making should follow ethical standards, taking into account environmental, social and governance considerations, which will prevent AI Investing in companies can be harmful to society.

In short, all of the above means that the AI ​​depicted in academic experiments is not feasible in the real world of the financial industry.

Are humans better?

If AI can make investment decisions as good as or better than humans, this means employment will drop significantly in the future.

In fact, some AI funds that publicly disclose their performance often underperform in the market, which further increases favor with analysts and human managers.

Empirical evidence shows that humans are now ahead of AI. This may be partly due to the fact that people often have a “gut feeling” when forced to make quick decisions in the face of uncertainty.

In the future, this may change, but we still need more considerations before relying on AI. However, instead of encouraging humans against AI, we should combine the two. This means using AI in decision support tools and analytics, but making the final investment decision should still require human intervention.

Huong Dung(According to TNW)

Robots serving cocktails at the 2022 Winter Olympics

Robots serving cocktails at the 2022 Winter Olympics

Organizers of the Beijing 2022 Winter Olympics have equipped a robotic bartender that can serve cocktails in just 90 seconds.


Leave a Reply

Your email address will not be published. Required fields are marked *