Five ways AI will (probably) affect the finance industry

Will AI augment human activity – or replace us altogether? Could governments manipulate or censor AI  in order to get them to surreptitiously do their bidding? What’s the best way to regulate AI  without stifling innovation? And finally: what does all of this mean for the finance industry?

These are just a few of the big thought-provoking questions on the minds of senior finance industry executives who joined a discussion hosted by RBS International on the role that AI will play in shaping the sector. While nobody truly knows exactly what shape it will ultimately take, consensus is starting to emerge among senior banking and investment management executives around areas where these technologies present clear benefits – and risks.

Machines can’t be trusted with empathy or relatability, so neither can relationship management

One fundamental question hovering above the heads of workers everywhere is: will I be replaced by AI?

The finance industry is not immune to the same fear, uncertainty, and doubt that accompanies innovation seen elsewhere in the economy. AI amplifies many of the pressures that digitisation and data-centricity create (skills and technology gaps) and trigger many of the same anxieties.

Yet it would be too simplistic to think that AI will simply replace finance sector workers en masse. People excel at two core qualities that appear impossible to ‘teach’ to machines: empathy, and relatability. Crucially, those qualities also underpin relationship management – an essential role in the finance sector, and one that doesn’t look likely to be handed over to the machines any time soon (no matter how compassionate they are trained to appear).

Automating boring, manual, process-driven (but ultimately very important) tasks

Rather than attempt to replicate the most ‘human’ of qualities, AI does offer very compelling benefits in supporting specific areas traditionally gummed up by bureaucracy: client onboarding; satisfying anti-money laundering (AML) and know-your-customer (KYC) regulatory requirements.

AI isn’t likely to replace the people ultimately accountable for these important functions. But it does have the potential to eradicate mundane identity checks and augment knowledge about customers.

The combination of Big Data and AI has huge potential, and the best is yet to come

AI may one day be able to help create a 360-degree view of customers, which may deliver benefits well beyond simply being able to satisfy compliance and regulatory requirements more cheaply and quickly or improving customer relationships.

A good example is connecting the fields of farmers to the world of finance: if we could monitor fields to see how crops are developing in real time, one might imagine a scenario where insurance could be taken out to mitigate the risk of poor crop yields. Equally, one may consider investing in a company based on wheat or barley yield successes.

There is also a huge opportunity here to use data and AI to make the world better, too. Combining AI with environmental, social and governance data (ESG) could help produce new and powerful operational tweaks that can help make business more sustainable, and help investment management companies identify ESG risks that may otherwise be nearly impossible to detect.

But where opportunity is found, risk usually lurks too. What happens if the data feeding these AI models are wrong? What if the models themselves are biased? The consequences seem trivial if, say, ChatGPT spits out a recommendation to dine at a restaurant that’s been shut for three months. They’re far greater when it comes to decisions governing how corporate balance sheets or billions of dollars in assets under management are put to work, or how operational tweaks are decided in order to improve decarbonisation over the long-term.

The growing impact of – and focus on – fraud

One only needs to watch a single deepfake video to grasp the potential impact that AI-generated content could have on financial institutions and their customers.

AI is helping fraudsters dupe people into making risky financial decisions, making them transfer money or open a bank account to launder money. But if you think it’s the simply digital novices or naïve at risk, think again. We’ve already seen one high-profile case involving the use of AI-based voice cloning software to impersonate a company director – who proceeded to convince a bank to arrange $35 million in fraudulent transfers from corporate accounts to private ones in another country.

There is a sense among finance industry leaders that, as technology improves, fraud will only get worse in the years to come. The financial sector will need to work hand-in-hand with customers and regulators to fight and prevent it.

More robust regulation is needed

AI offers the potential to make the world a better place. But, clearly, it also confronts us with new risks and challenges. These raise important questions about how regulators best treat AI: how should regulators treat the ethics of AI? How can we ensure AI doesn’t evolve in ways that bolster criminal behaviour (and get better at masking it)? How should regulators deal with accountability and legal recourse in an environment where they are making financial or investment recommendations to customers?

Although the answers to these questions are by no means clear, one thing most can agree on is the need for careful oversight from regulators. AI technology is evolving and new applications proliferating faster than regulators can move and businesses can develop effective and appropriate governance structures. Regulation must be carefully considered if it is to effectively guarantee security, safety, and privacy without stifling innovation.

The road ahead

Risks and challenges notwithstanding, AI is an extremely exciting space that leaves many senior leaders in the finance industry optimistic about the future. Change and disruption is undoubtedly on the way, but it seems clear that the potential to transform the industry – and our lives – in positive ways will only grow.