The world has been caught by storm with the newest generation of AI technology. Functions like ChatGPT and Bard have shown exciting and terrifying new prospects for nearly every single sector of the labour market, from legal and financial services to the arts and media. Investment banking is a sector, where it is expected that AI technology will play a pivotal role, whether speeding up admin and data analysis work or automating trading capabilities. This article will look into what the current development of AI will change in the future of finance, looking at its uses both current and future, and understanding the risks that this new technology presents.
What is AI currently used for in finance?
The world has become a data driven place, especially with each new technological revolution that humanity experiences. It is estimated that in 2020, every human on average created at least 1.7 megabytes of data every second. The finance world is a very data-driven sector, meaning AI has become a useful tool to automate a huge number of tasks, often ones that are tedious or can have a large degree of human error.
One predominant use is in risk management. The financial markets can be a risky and unpredictable place, carrying worries of handling in bad faith, incurred losses, and human error. AI finds two uses in this spheres, due to its expert ability to identify pattens and correlations. Firstly, it can quickly identify irregularities and unusual trading patterns. This quickly allows the programme to identify potential market manipulations and fraudulent activities with an efficiency that humans simply are not capable of. As well as this, it is able to use its pattern seeking abilities to push much greater predictability in an extremely efficient manner, improving the quality of data, and eliminating the effect of human risk.
Another use is data analysis. AI can analyse vast amounts of data, much more and much quicker than a human stock trader, and is able to quickly identify patterns, correlations and provide insights for better investing. It has an incredibly low margin of error, allowing investors to make factual, data-driven decisions, improving the quality and efficiency of their work.
Large companies such as BlackRock and JP Morgan have begun exercising reliance on AI technologies, with BlackRock for instance using self-learning AI code in its exchange-traded funds business, reducing its reliance on human stock pickers. Thus, AI firmly has got its grips in the investment world. It cuts down risks, increasing data analysis output and provides a greater data quality.
What has changed?
What has changed in the realm of artificial intelligence is the introduction of generative AI. Up until now most AI systems including those discussed above fall into the category of predictive AI, a form of technology which uses patterns in historical data to forecast future outcomes or classify future events. This is especially useful in analysing market trends and conducting data analysis. Generative AI on the other hand is a novel version, which focuses on creating new, creative outputs that mimic human-like patterns. When looking specifically at the banking world, the most common gen AI that is coming into force are large language models (LLM). These are artificial intelligence models that have been trained on vast amounts of text to understand existing content and generate original content.
These new systems are poised to completely reform the world of investment banking and increasing productivity, with a McKinsey report explaining that banks and the finance sector are expected to see increased productivity of 2.8-4.7% in the industry’s annual revenue or an additional $200 – 340 billion. For front-office operations, gen AI could increase productivity by up to 35% by 2026, adjusting for inflation.
The possible uses of gen AI in investment banking are countless. One example is in customer service capabilities. A lot of banking requires B2C, direct human interaction. Naturally, this currently means specific calling hours and huge numbers of call centre employees to deal with the huge numbers of phone calls. A LLM gen AI bot, however, if it is trained on all of the required knowledge, including policies and research, can provide good, in-depth customer information. This service can be 24/7 and can save banks huge amounts of money. One European bank, for instance, is implementing an AI virtual expert specialising in ESG (environmental, social, governance) investing, allowing the bot to tackle an extremely complex topic in a hyper-efficient manner.
Are there risks involved?
One of the big questions involved in AI analysis is the risks thereof. Many have an image of AI in their heads, which is most likely inspired by 80s dystopian sci-fi tropes. Thankfully, this isn’t the world we live in yet and hopefully it won’t ever be so. However, one point that Harrison Ford and Arnold Schwarzenegger do make which is applicable in any AI situation is the necessity of a duty of care, with there still being a number of risks that can arise in a financial context.
One concept that has been talked about a lot recently is the idea of impaired fairness. This means the AI projecting algorithmic bias due to issues in the training or engineering stages. Decisions that are made with AI deal with real life situations and affect real life people, so any biases present within the algorithm can exacerbate social inequalities and the vulnerability of minorities. One method to mitigate this is keeping a balance between human expertise alongside AI. This means that AI decisions don’t go unchecked and thus can attempt to reduce possible biases or discrimination.
General AI is a very novel technology, and as such many new applications still have discrepancies and bugs throughout their algorithms. This can promote grave security concerns, with users or hackers being able to manipulate the system to their benefit. There are also privacy concerns involved, through the unintended use of client-sensitive information in model training.
AI does present a scary new reality that many of us can’t truly grasp. Nevertheless, this risks are ones that, with an adequate duty of care, can be successfully managed and mitigated.
To conclude, gen AI presents a fascinating new frontier in investment banking. It will bring billions into the sector, and promote a brand-new employment sector. It will bring a new improvement in efficiency that hasn’t been seen since the internet Revolution of the late nineties. While there are risks that need to be properly managed, the benefits of gen AI will bring out wonders for the investment banking world.