Introduction

Technology has been spreading its influence around the globe, becoming a truly substantial part of daily life, changing many jobs and perspectives on what can be done. As this golden age of technology continues to spread its influence, Artificial Intelligence (AI) has undoubtedly become a very big part of its discussion, a tool that surpasses human intelligence, through its efficient method of data analysis. As an important indicator of the U.S. economy, the stock market has always been difficult and complex to truly predict. AI has become a very controversial topic amongst investors, and many wonder about the impact AI will have on the field of financial analysis, and the limitations it has.

Many different studies have found ways AI can be used to benefit from its ability to scan through data. Backtesting, a process in which previous data from history is used to see how a contemporary strategy would perform if employed in the past is significantly improved by using AI's ability to search through data. AI has slowly become a fundamental factor in many parts of data analysis, used by many quantitative investment firms hoping to uncover hidden patterns and trends that would have been otherwise missed due to human error. Although the benefits of incorporating AI seem plentiful, the obstacles in its way are truly substantial. Nicolas Taleb in Taleb (2007) found that throughout the course of history, there have been many events that are almost impossible to predict called: “black swans”. These black swans are an insurmountable obstacle to AI when it slowly becomes incorporated into the field of financial analysis. Overall, from examining the state of AI in financial analysis, there is enough evidence to say that although artificial intelligence provides us with many benefits, we cannot ignore the challenges when implementing AI in financial analysis. This research paper will attempt to explore all the opportunities further, while also looking at the challenges and their limitations.

1 Backtesting With Artificial Intelligence
Studies by Jiarui Ni and Chengqi Zhang from the University of Technology, Sydney, found that when implementing AI algorithms into backtesting strategies, the time it took for the algorithm to provide results was too long. Therefore, it would make that method obsolete because users would not wait that long for results. To combat the length of the processing for results, they found that parallelizing its executions is more efficient and less time-consuming. However, with that method, it would require a mass amount of computational power, due to the number of processing elements (PE) needed. Users can not be expected to have a computer good enough for the requirements of the computing time in order to receive results. Finally, the results showed that Genetic Algorithms are much more efficient and have more capability when it comes to handling a large amount of data. With a Partheno Genetic Algorithm (PGA), they achieved results within 1 to 2 hours, greatly shortening the time compared to the two other methods. A genetic algorithm is derived from biology and uses the terms generation, mutation, parents, and offspring. A programmer is to input a given population and the best two individuals (parents) are to make offspring-generation 1. The offspring(s) then undergo mutation, and the process repeats until we achieve a desired result or a specific number of generations. This method drastically reduces both computation time and power. 

2 Smart Beta
Smart beta consists of active and passive investing, becoming one of the most talked-about and popular strategies. As the market is becoming riskier and volatility increases, investors have relied on smart beta and its safety. This popular strategy involves using an algorithm to pick stocks in an extensive index. Unlike an alpha strategy which is risky and expensive, a smart beta strategy is passive and moves with the market in order to receive returns. According to Ung, Daniel, and Priscilla Luk (2006), “ smart beta strategies can refer to a swath of strategies that are designed to provide access to a wide array of return-enhancing risk premia (or risk factors)”. Basically, smart beta focuses on the potential factors in a transparent way.

3 Defining a Black Swan 
A black swan event, defined by Taleb (2007) is the impact of the highly probable, unpredicted event contributing to the cause of many cases of misleading patterns in data. Taleb (2007) argues that “black swans” tend to consist of three characteristics: 1. They carry a substantial impact; 2. They cannot be predicted due to their extreme rarity; 3. After a black swan event happens, the event seems obvious in hindsight. These black swans are found in various points of our past, defining many events that have changed the course of history. A black swan can be found to either carry positive events (e.g., world-changing inventions that seemed to come to fruition by luck) or negative events (e.g., a country’s unprecedented attack on its neighbor). One of the most dangerous and deadly black swans is the coronavirus (COVID-19) in 2020. With the rise of this global pandemic, the market has reacted with a major crash, causing many stocks to plummet.

4 Artificial Intillegence’s Benefits and Limitations
The factors that have been included so far all fall under the umbrella of financial analysis, but there are many benefits and limitations when it comes to AI in this field. Backtesting is one of the most important strategies for all investors and their companies. The importance of backtesting to a trading system is clear, and many contemporary algorithms have found ways to speed up the results of a backtest. With these AI algorithms, we have found a way to reduce computation time tenfold, and it only gets shorter as we explore the vast possibilities of AI. Smart beta has been a controversial topic, on its way to achieving alpha. Investors have found the smart beta to be an efficient way to find multiple investment strategies. Aside from the benefits, a major limitation of AI in the field of financial analysis and market prediction is Nicholas Taleb’s “black swans;’’. These swans are almost impossible to predict and can influence the market greatly. AI cannot yet realize the dangers of these swans, and there is no solution to solve them. The truth is that, unlike AI, we can realize the situations of the world of black swans, and try to redirect these events into positive outcomes.
5 Conclusion 
This paper aimed to put together possible benefits and limitations when incorporating or even investing in AI and it can be seen that the benefits outweigh the side limitations. AI is one of the most important factors in contemporary trading strategies, but when it comes to market prediction, human intelligence is needed. Without the implication of artificial intelligence in backtesting and the genetic algorithms we have developed, the calculation time would make backtesting almost unviable. Without the implication of artificial intelligence in smart beta, we would be forced to take bigger risks and make it difficult to safely receive returns. Overall, artificial intelligence has made a substantial contribution to our stock and business industry, and as it continues to be studied, it will surely continue to further improve the way we invest in the market.

Works Cited

Ung, Daniel, and Priscilla Luk. “What Is in Your Smart Beta Portfolio? A Fundamental and Macroeconomic Analysis.” The Journal of Index Investing, vol. 7, no. 1, 31 May 2016, pp. 49–77, 10.3905/jii.2016.7.1.049. Accessed 23 Nov 2022. 

“Backtesting Comparison of Machine Learning Methods on Warsaw Stock Exchange.” Elsevier Enhanced Reader, 2021reader.elsevier.com/reader/sd/pii/S187705092101886X?token=8E0D863FA9A02D4619F824F6B4F0E0BA03CD37F671FF0BB7749D4455B821D151BC1F766AFEDDB13D0DA0D3DE8FF3B7DF&originRegion=us-east-1&originCreation=20220516025808. Accessed 23 Nov 2022. 

Rodriguez, Jesus. “The Black Swan Problem in Artificial Intelligence: Part I.” Medium, Medium, 20 Mar. 2017, Accessed 23 Nov 2022. 2019jrodthoughts.medium.com/the-black-swan-problem-in-artificial-intelligence-part-i-74306aee0156#:~:text=Black%20Swans%2C%20by%20definition%2C%20are,yet%20materialized%20in%20AI%20models.

Ni , Jiarui, and Chengqi Zhang. An Efficient Implementation of the Backtesting of Trading Strategies. 2007, web.ist.utl.pt/~adriano.simoes/tese/referencias/Papers%20-%20Pedro/Backtesting.pdf.

S&P 500 Forecast 2021 – Long-Term Prediction – Outlook, kagels-trading.com/sp500-price-forecast-longterm-prediction/. Accessed 15 May 2022.