Tell your machines not to mess with the pyramid, it’s important
GlobalCapital, is part of the Delinian Group, DELINIAN (GLOBALCAPITAL) LIMITED, 4 Bouverie Street, London, EC4Y 8AX, Registered in England & Wales, Company number 15236213
Copyright © DELINIAN (GLOBALCAPITAL) LIMITED and its affiliated companies 2024

Accessibility | Terms of Use | Privacy Policy | Modern Slavery Statement
People and MarketsCommentGC View

Tell your machines not to mess with the pyramid, it’s important

pyramid.jpg

Investment banks should not start replacing entry level positions with AI-trained computers, it could undermine the whole system

Just this week a photograph created by artificial intelligence won a prestigious competition, and a song generated to sound like famous pop stars had the music industry reeling. AI and its use cases are advancing, as are age old fears that the technology can replace humans in not only mundane, but also specialist tasks. Tech leaders around the world, including the ever-present Elon Musk, have called for a slowdown in development while we figure out exactly what it is we are doing.

Indeed, GlobalCapital wrote that AI tech like ChatGPT is more likely to improve investment banking jobs than replace them. But it is worth considering the long-term impact that this sort of technology might have on the capital markets.

Every year, thousands of trainees and interns start careers in the capital markets. Goldman Sachs received more than 260,000 applications for 3,000 internships in 2022, an acceptance rate of just over 1%. Graduate training schemes and internships have long been an important step in the career development of anyone looking to enter the industry.

The grunt work at that level can be gruelling and may extend to such illustrious tasks as adding logos to endless presentations, plugging numbers into spreadsheets, or fetching the sandwiches for lunch. It is hardly the stuff of the Wall Street big swingers but the the skills learned and aptitude — not to mention attitude — demonstrated at this stage are crucial for juniors' development.

Meanwhile, junior hiring is on the slide. Deutsche Bank said that it took on 11% fewer graduates in 2022 than the previous year, and while headcount remained high (thanks in part to many hires in India) this is likely indicative of how things have played out across the Street.

With advances in AI and pressures on costs and return on investment it will be very tempting for bank management to look to replace juniors with AI or other tech.

But there are good reasons to keep the junior ranks of the front office well stocked.

Firstly, as in any industry, a pyramid system of seniority works. There is more legwork to be done at the junior level and those that show they can do it are those who will demonstrate the skills, aptitude and who will develop the trust and experience to rise into senior positions. Inevitably, some will have made the wrong career choice but people develop at different speeds and they and their employers need to go through that sorting process while the bank has enough staff to do the work.

It isn't that banks cannot be made more efficient but there is also value in juniors doing the donkey work in the early stages of one's career — recording each new issue, for example, to be able to spot patterns and to develop a sense of the story of a market. And, frankly, how can you be expected to run an order book if you can't get the lunch order right for your bosses and deliver the correct change (though it has been said that those that manage to keep the change are ear-marked for even greater success)?

Increasing margins by hiring fewer juniors also reduces the pool of talent for the future. When the current and next crop of MDs quit the business, there is a risk that those left to be promoted will not have had the deep learning experience of the junior years if a lot of the data crunching and so on has been done by machines rather than by an analyst or associate in Excel.

Secondly, although AI moves fast, much of the valuable work in capital markets cannot be automated. AI cannot generate a list of appropriate comps to show a new issue in its best light, for example. That requires the input of experienced traders who know the subtleties of their market and someone with an understanding of the nuances of new issue process.

Capital markets are at their heart about human relationships and trust based on experience and expertise; art and science. That is really what is being developed among the junior ranks. That may be a hard slog at first and it may involve some unpleasant moments. But a junior will learn swiftly from a rebuke from a manager. Shouting at AI has yet to be shown to generate fast improvements.

Like much of the world, investment banking is staring down the barrel of an AI shaped gun and working out how best to proceed. It will have its uses but they won't be half as numerous as the benefit of bringing on subsequent generations of humans, at least not until ChatGPT figures out how to get your lunch to your desk.

Gift this article