Katharine Wooller, Chief Strategist – Financial Services at Softcat plc examines AI having the potential to democratise the finance industry
Evangelical supporters of AI see the technology as enabling a new frontier of progress for mankind; they breathlessly expound its virtues, and suggest, that by expanding our ability to process data, we will be able to unlock hitherto unachievable levels of progress.
Top prizes on offer include a cure of cancer, and the creation of financial and educational equality. The more realistic amongst us roll our eyes – having heard the same wild promises from proponents of blockchain, who were also adamant that a new methodology for data was just the ticket to achieve the same lofty achievements.
The jury is out, however, on how and if AI will create financial equality – with the whopping great caveat, if it is designed and deployed responsibly.
There are some persuasive arguments to suggest that AI could power a much fairer society. For the first time, in theory, deploying large language modules that are free to access, will provide swathes of people with access to quality financial education. Whilst I am not suggesting that the “man on the street” has access to the algo magic available to hedge funds, there should be a democratisation who can find, for example, detailed analysis of stocks and shares.
There are swathes of the population who are excluded from banking services and particularly credit. With AI-driven credit scoring and AML derived from new and increasingly sophisticated data sets, those with patchy credit history should be in a measurably better position. Similarly, much work is being done, particularly by the neo-banks who are digital native, to give genuinely personalized financial coaching, providing free, tailored advice – for example on budgeting, savings and investment advice appropriate to their stage of life and wealth bracket.
Ultimately the drive to AI technologies is driven by the desire to reduce cost and risk; in theory this is good for the consumer. There are some extremely positive use cases for AI finding fraud in real time (for example in push scams) which protects vulnerable users of the banking system.
By increasing automation, in theory, banks should be able to lower the cost of providing banking services, although quite how much of this cost saving is passed on to the end consumer remains to be seen, as the landscape in retail financial services is hugely competitive already, and margins tights.
Again, with increasing efficiency in the administration of financial services, you would hope to see this filtering down to more inclusive opportunities for previously under-served sections of society, which was a strong argument for peer-to-peer lending (and the microloans it was able to provide) a decade ago, and for crypto, more recently.
However, as with all technologies, the early phases of adoption can see some inherent dangers, and many examples of “unintended consequences”. Left unchecked a new technology can be potently dangerous, and there are some examples where AI is deepening the wealth divide. Most obviously there is a huge variety in how governments internationally are fostering innovation.
Due to the huge amount of computational energy and infrastructure needed to run large language models there an “arms race” by those countries looking to be a competitive home for leviathan tech companies that will generate huge amounts of tax revenue and jobs. Those nations, and their citizens, unable to keep pace risk being left behind.
Without checks and balances, there is significant evidence that AI could even worsen financial inequality. There is much concern over the risk factors of AI becoming biased or hallucinating, which can lead to discriminatory algorithms – this is a huge problem if you are using AI to work out who to allow to vote, or get a loan, or for recruitment.
There will doubtless be significant social disruption by the job displacement in low- and middle-income jobs as their roles are replaced by AI driven automation and agents. Without significant learning and development to keep workers’ skills relevant there is likely to a large digital and earnings divide for those who do not understand, and embrace, AI.
There is much concern over the concentration of wealth and power with the tech giants who control powerful AI models, which uses the data of their users to create LLMs; Microsoft, Meta and Google are already some of the most valuable companies on the planet, which seems far from equitable.
In short: AI has the potential to democratize the finance industry — but without regulation, transparency, and equitable access, it risks supercharging existing inequalities. It will be interesting to see how regulators worldwide, whose primary objectives are to create financial stability and treat customers fairly, grapple with this issue for years to come.
