Over the past few months, I’ve had occasion to speak at a number of conferences concerned with the impact of artificial intelligence on financial jobs. My audience’s interests vary. Some are students and young professionals concerned about job prospects. Some are bank executives and investors interested in employee cost projections. Some are customers and regulators who want to know how AI will change client experiences and protections. And in some cases, they are interested in learning how to exploit opportunities to sell to established financial institutions or to compete with them.
I have one answer for all of them. AI advances of the last five years or so will completely eliminate some large categories of financial jobs that have been around for many decades. However, like financial innovations of the past, it will also create jobs in two different ways. The improved efficiency and power of AI methods will create entirely new opportunities, and the improved performance of financial systems will stimulate growth in the sector.
But even if the net result is more and better jobs for humans, there will be massive disruption. The people who get the new jobs will not necessarily, or even usually, be the people laid off from the old ones; and the companies exploiting the growth in the financial sector may be recent start-ups rather than established institutions.
The message for financial students and workers is to prepare for change, the message for investors is to diversify among existing major financial institutions and promising innovators. Customers and regulators should anticipate fundamental changes in their relationships with financial institutions, and people looking for opportunities should prepare for a gold rush.
Because both AI and financial jobs are big categories, it’s easy for discussion of the impact to lose focus. I’ll start with a specific version, a brief history of “know your customer.” Traditionally, this meant a bank officer literally knew the customer, including his family, business and business associates.
In 1970, the US Bank Secrecy Act and similar legislation in other jurisdictions created the first specialized KYC jobs. There was little automation of any kind. Junior personnel verified names and addresses of customers, paid visits to client businesses and checked information against official records. Over time, more of this was computerized, but growth in the size and complexity of the financial sector, plus increasing regulatory demands, meant KYC workforces grew despite technology making each worker more efficient.
In the 1990s, international bodies such as the Bank for International Settlements and Financial Action Task Force were the drivers of regulation, with global standards replacing patchwork national rules. KYC changed from documenting basic customer information to ensuring complex due diligence procedures had been carried out — the focus shifted from the product to the process. This required more automation, and hiring more specialized KYC personnel, including more higher-ranking and higher-paid employees.
In the 2000s, the US Patriot Act turned KYC from a monitoring activity designed to protect the bank from credit losses and lawsuits by excluding shady customers, to a major arm of the government’s law enforcement, tax collection, terrorist fighting, corruption detection and sanctions enforcement operation. This vastly increased the necessary information collection and processing, and the numbers of KYC specialists, who now included lawyers, information technologists and analysts. KYC went from some paperwork hurdles needed to open an account, to ongoing monitoring and deep research. KYC now employs something like 10% to 15% of bank workers.
Nearly all of this will soon be done by AI. A relatively small number of technologists — AI specialists, not KYC specialists — will be needed, along with a few very senior people to set policy and investigate major cases. The armies of investigators can be replaced by algorithms running 24x7, collating customer information with all kinds of structured and unstructured data on the internet, and deploying the understanding of the world that AI has developed in recent years.
No doubt there will be an arms race, with people exploiting AI to fool the bank algorithms. But I expect this battle will be joined by government security agencies rather than governments contracting the work out to individual banks. This will increase demand for crime fighters, tax collectors, anti-terrorism people, anti-corruption regulators and sanctions enforcers, along with data scientists and IT developers, not bankers.
What will happen to the hundreds of thousands of bank employees working in KYC? I hope they all have backup plans. The AI technology implemented to perform KYC is likely to open up more opportunities than it destroys. After all, knowing your customer does more than exclude some dicey counterparties, it helps you serve them better, and to find new customers.
While details vary, the basic KYC pattern will be repeated in other bank departments such as operations, compliance and risk management. A human activity — bank officers knowing their customers — is replaced by a manual process for more junior people. Computerization makes the process more efficient, but growth in finance and increasing demands for information keep the head count growing. Rote tasks are delegated to machines, which are faster, cheaper and more accurate, but humans are needed for their knowledge of the world — for knowing that “Richard” and “Dick” can be the same name, and that there are no cattle ranches in Antarctica, for viewing with suspicion a low-level government employee asking for a loan to buy a yacht. Recent advances in AI mean computers are now much faster, cheaper and more accurate than humans for these knowledge-of-the-world tasks as well.
The financial system of 2030 will create much more economic value than the system of 2025, and will offer much more opportunity for good, high-paying jobs. But most of the 2030 jobs will be quite different from 2025 jobs. Much of the 2025 economic value of finance will be performed by machines, and the successful 2030 human financial workers will be the ones who exploit the power of AI, not the John Henrys who try to compete with it.
A message from Advisor Perspectives and VettaFi: To learn more about this and other topics, check out some of our webcasts.
Bloomberg News provided this article. For more articles like this please visit
bloomberg.com.
More Innovative ETFs Topics >