By Pia Kothari
Unlimited access to open-source artificial intelligence software allows users across the face of the internet to simplify commercial and non-commercial activities with greater efficiency and speed than most traditional methods. The evolution and subsequent development of artificial intelligence date back to the mid-20th century and has since flourished its way into the 21st century. Artificial intelligence, however unconventional and convenient, has its potential disadvantages for society. For this, it has caught the attention of many analysts and researchers, mainly investigating the “grave” dangers of its excessive use and dependence.
Focusing on its danger to job markets, predictive and hybrid artificial intelligence software paired with neural networks actively poses as a threat to jobseekers in multi-disciplines and is forecasted to absorb over 90% of jobs over the next 10 years. This has a significant impact on the world economy and may increase international income inequality. Financial management professions tend to be one of the most threatened by AI wherein human capital may be made partially or completely redundant.
Ranging from the formation of indexes to the creation of arbitrage strategies, AI has changed the face of the financial management industry. However, the accuracy and ability of several AI platforms like AlphaSense, QuantConnect, Kensho, EdioSearch, Ayasdi and many more remain widely debated upon. Often used together, these AI-powered platforms execute algorithmic trading and investment strategies powered by machine learning and unique quantitative finance techniques. They seek to make fast-paced decisions and calculated risks tailored to fit the profile of every user to take advantage of any opportunity in the market. Common users include MNCs, Hedge Funds, Alternative Investment Funds, LBO Funds, Mutual Funds, REITs, Pension Funds, Commercial banks, individuals and ultimately any market participant. It is to be noted that some of this software is privately owned and patented. In addition to leveraging investment positions, the trades made by these AI platforms are executed in a matter of seconds, almost tracking price movements with agility, beyond the comprehensive knowledge of traditional white-collar analysts. Although agility of an AI system may be unbeatable by human cognition, the accuracy of these AI powered investment recommendation systems seems to be in the grey. The question remains: Can Predictive AI completely replace buy-side analysts?
Even though buy-side analysts are time bound, agility is a seemingly rigorous requirement for buy-side traders, tasked with the sole responsibility of execution of transactions and investment strategies of money management firms. The argument continues about AI regarding its lack of direct human creativity in its formation of investment choices and strategies persists among the asset management and banking industry. Most predictive AI models use machine learning to analyse past events, market sentiment, fundamental data, etc. and accordingly make their recommendations and investment strategies. With a successful integration of behavioural and trend analysis, the model may provide near-accurate predictions, however, the problem lies with its lack of ingenuity in combining their results to make an efficient investment strategy with an effective execution layout.
To discuss further, buy-side analysts bring a significant amount of human judgement and industry expertise to the table when making their recommendations to fit a client’s portfolio. Many are concerned that AI, as promising as it seems, doesn’t have a “human touch” to it. Not only does AI lack the required innovation and adaptability to strive in the financial management industry, it also doesn’t have a face to it. In such an industry, it is almost a requirement to have a “trustworthy face”, one which people could entrust their money with. This builds healthy client relationships with buy-side analysts, who usually tend to work on the back end of a firm, nevertheless, are required to interact with their clients.
Highlighting this fact, many debate otherwise that AI makes the work of these analysts simpler, if not replacing them, by its powerful ability to condense, interpret and present large amounts of unstructured and structured data, which isn’t only constrained to numerical data and text analytics. Based on its data analysis, AI enables analysts to recommend tailored financial products and make investment decisions and strategies at a pace faster than ever before. In a recent development, BlackRock introduced its AI platform, Aladdin, which seeks to provide comprehensive risk assessment and portfolio management solutions. If used in collaboration with human judgement, “AI has a huge potential to increase productivity, increase knowledge base and transform margins across sectors” said BlackRock CEO Larry Fink in a recent interview. Similarly, in order to improve trade execution and efficiency, JP Morgan introduced the LOXM programme, which leverages AI to enhance the execution quality of their trading strategies and minimise the market impact. Interestingly, the LOXM programme is powered by machine learning and effectively performs predictive risk management and scenario analysis using historical trade data.
Aside from assisting and improving the primary duties of any analyst in the financial industry, AI also enhances the operational capabilities of their firms. As an example, UBS employed AI algorithms to provide enhanced CRM functions through their SmartWealth platform, which provides personalised investment advice after considering individual client profiles and risk appetites. Proven to be cost efficient for a labour intensive financial services firm, AI streamlines back-office operations in order to improve delivery and efficiency. This is observed in Goldman Sachs, which utilises AI to automate complex and labour intensive tasks and processes, aiming to reduce costly errors and save time. Referring to costly errors, AI also aids in fraud detection, most prominently used in Citigroup as a measure to identify unusual transaction patterns and provide real time alerts.
In the hope of utilising AI efficiently as an active cost-cutting measure, firms must evaluate technical, legal, and commercial considerations before committing to complex AI models in their analytical and operational activities. Nevertheless, most AI algorithms may inherit biases present in the data due to their machine learning property, which could lead to potential errors in their analysis. As a preventive measure, AI systems must be regularly monitored and adjusted to avoid any erroneous analysis and interpretation. Like all technology, concerns were raised regarding cyber security of sensitive financial information stored in the memory of AI systems and automation technology. Data of high value may be at significant risk due to heavy dependence on digital systems and lack of cyber security measures.
Ultimately, the debate continues to grow and become stronger as and when new advancements are made to existing models and algorithms of AI or the development of new systems and platforms. Many experts and computer scientists believe that Artificial Intelligence is still in its early stages of development. Even though it has already touched the lives of several individuals around the world, many do not know its capabilities and functions to use it efficiently and effectively. In the near-future, with respect to its application in the financial management industry and the possible replacement of buy-side analysts, it is highly unlikely that these AI systems will be able to fill the shoes of buy-side analysts in an investment management setting due to its inability to fully capture and analyse qualitative information, human intuition, and market sentiment. However, the future remains uncertain concerning the potential of Artificial Intelligence.