Artificial intelligence (AI) is automating many entry-level white-collar jobs in financial services and other industries, according to data analytics firm, CB Insights.

Banking, FinTech, and insurance are among the sectors being most disrupted by the influx of the technology, alongside legal services, marketing, industrials, and energy. AI’s ability to identify patterns within masses of data could give rise to critical insights.

The company’s findings underscore those of the World Economic Forum last year that AI, robotics, automation, and other so-called Industry 4.0 technologies will decimate once-safe middle-class careers.

The WEF report, The Future of Jobs 2018, forecast that 75 million jobs worldwide will be displaced by the technologies, but said that up to 133 million new human jobs may be created by them: a net gain of 58 million.

Despite this, the WEF painted a bleak future for roles that were once considered open doors into lifelong careers. Financial analysts, lawyers, accountants, auditors, bank tellers, general managers, business services managers, and administrators were all listed under “redundant roles” – over a timescale as short as the next five years.

According to the WEF, among the jobs taking their place will be data analysts, AI and machine learning specialists, innovation and digital transformation experts, robotics professionals, user experience and interaction designers, process automation experts, and other creative roles that machines may find hard to replicate – for now, at least.

Interest in AI is certainly soaring, according to CB Insights’ figures: for example, since 2016, over 1,100 AI startups have raised their first investment rounds. In 2017, equity funding of the technology hit over $15 billion, a year-on-year increase of 141 percent, and three times more AI companies – over 300, in fact – entered incubators.

At the core of this surge is a group of technologies known as deep learning, which Google and others are heavily investing in. In many ways, deep learning is designed to model the workings of the human brain, identifying and classifying images and text and understanding the nuances of language.

In financial services, its applications are legion. They include: fraud prevention; predictive analytics; trading; regulatory compliance (RegTech); investment; capital markets operations; claims processing; and even customer service – indicating that the technology is encroaching deeper into human territory, initially as a competitive differentiator.

As far back as 2016, Bank of America began trialling an automated investment platform at its Merrill Edge subsidiary. Last year, HSBC invested in AI-based fraud detection systems, while NatWest claimed to have prevented £7 million of corporate fraud using machine learning.

Danske Bank is another big name deploying machine learning to detect fraud in banking and mobile payment transactions – but the technology may also be being used by fraudsters and money launderers.

Last year, Danske’s chief executive fell on his sword when it was discovered that the bank’s Estonian branch had been used to launder over €200 billion (£180 billion) over a nine-year period – potentially the biggest such scandal in history. Banks are fighting fire with fire.

Meanwhile, Lloyds Banking Group partnered with anti-fraud specialist Pindrop to identify fraudulent phone calls, India’s ICICI Bank deployed AI for sentiment mapping in customer contacts, and DBS, Barclays, Wells Fargo, Santander, Caixabank, and Swedbank all implemented AI in customer service support.

Some of these may be low-level implementations, such as chatbots, but that indicates that AI may begin to sweep aside low-skill, high-churn roles, such as call centre operators, too. In the UK, up to four percent of the entire workforce is estimated to work in some 6,200 call centres – often in areas where manufacturing and other traditional industries have declined.

With Brexit looming and a broad range of industries, including banking and manufacturing, considering their options and inward investments, this could pose a genuine threat to long-term UK employment.

Also in 2018, the Bank of England began testing technology from MindBridge to spot abnormalities in financial transactions, while Bank of America, ANZ Bank, and JPMorgan, among others, deployed AI in contract analysis and loan approvals.

But such moves are not without their risks. Should AI, machine learning, and deep learning be used to deny some people finance or insurance services, for example, banks will have to prove to regulators that they are not discriminating against legitimate customers and that suspicious patterns of behaviour constitute hard evidence of crimes.

The risk of institutional bias in training data is real, and banks will be watched to ensure that they are not using AI to give discrimination a veneer of data-evidenced acceptability.

Other disruptive technologies are on the rise in banking and financial services. For example, 75 banks joined the Interbank Information Network last year, an experimental blockchain project to explore whether distributed ledger systems can speed up payments that have been held up by compliance checks or missing data. Also in 2018, IBM launched a blockchain-based clearing network for banks.

According to CB Insights, hedge funds and FinTech companies are even partnering with AI startups that analyse satellite images using deep learning in order to track oil, crop, and water inventories, programmes that may also benefit from distributed ledgers as a system of record.

Smart supply chains and payments may soon be part of the same, end-to-end, integrated system, which is already being used for carbon offset trading, for example. But it must be said that if it can be used to detect fraud, then AI could also be used to hide and automate frauds that may become too complex for humans to trace.

Despite all this investment activity in AI, however, banks and financial institutions themselves are losing ground in deep learning research, claims CB Insights – with no sector-specific patents in their name.

The risk of falling behind the technology curve in this sector can be given a name: Amazon, which is increasingly moving into financial services and payments. For example, Amazon India recently unveiled its Seller Lending Network, which uses machine learning to facilitate loans.

In China, Alibaba, Tencent, Baidu, and Ant Financial are all on the rise, mirroring the ascent of Amazon, Google, Apple, and others in the West, in offering AI and payment services. Few would bet against Facebook in this market either.

Inevitably, this means that the war for talent between the old financial guard and so-called ‘bigtech’ is increasing. In 2018, Goldman Sachs poached Amazon Web Services’ Charles Elkan as MD of machine learning and AI strategies, while JP Morgan employed Carnegie-Mellon University machine learning professor Manuela Veloso as head of AI research.

But one thing should be clear: with Industry 4.0 technologies kicking out the first rungs from the career ladder for many would-be professionals, a gulf is beginning to open up between newly qualified professionals or young trainees and their opportunities to progress.

In many cases, they would be better off learning to code and analyse data than learning to deal with customers and money – a real challenge for this customer-service industry, and others like it, such as legal services.

Be part of a discussion and connect with like-minded leaders in your sector at our exclusive event series on banking and RegTech.