Artificial Intelligence (AI) has been on the world’s tech radar for more than 60 years, but it’s never been more visible, or accessible, than it is right now. From machine learning and quantum computing all the way down to chatbots, AI is quickly becoming reality for organisations across industries and time zones.
By 2020, AI is expected to be a $5.05 Billion market. By 2035, experts predict it will double economic growth rates in 20 countries, and boost labour productivity by a whopping 40 percent. And let’s not forget the major disruption AI is already bringing to once-traditional industries like healthcare. Many experts say that the financial industry is next.
AI has steadily gained popularity in financial sectors because of its ability to streamline processes and tailor information to customers’ unique needs. We’re now seeing insurance agencies use AI to fight fraud, and financial advisors begin to explore robo-advisors. Banks are tapping AI to generate financial reports and process massive amounts of data faster—and with a higher level of accuracy.
But with any emerging technology, AI comes with risks. In finance and other sectors, many experts are still hesitant to put too much power into the hands of machines. Why? Because despite the amazing capabilities of machines to engage in deep learning and program themselves to accomplish what was previously assigned to humans, most people don’t actually understand how that works.
For banks and financial institutions, the ability to maintain compliance to regulations is another crucial piece of the pie. Most are in the earliest stages of AI implementation, enlisting robo-advisors to provide portfolio selections, and execute customer decisions. Only 32% are utilizing things like predictive analysis, recommendations and voice recognition and response. Those in the intermediate stage often rely on machines to assist customers with their goal-based planning, and recommendations. Advanced stage AI implementation in the financial sector requires in-depth understanding of customers and their finances to recommend and oversee their goals and portfolios. But trusting machines to readily give information and advice based on algorithms and recognized user patterns comes with a steep price tag. Without an understanding of how these machines learn or come to conclusions, the results can be disastrous.
The time to understand and implement AI in the financial sector is upon us. Here’s a quick primer for financial services professionals on the history, opportunity and challenges of AI.
The rise of AI in professional services industries
AI is far from a new concept. These ideas have been around in varying degrees since the 1950s, but really started to take off in the 90s with the automated digitization of handwriting. Over the last few decades, innovators have developed concepts like deep learning, which is the predominant AI outgrowth we’re seeing today. With the inception of chatbots and new smart machines, it was predicted that by 2010 people would start to depend heavily on machines and virtual assistants, and by 2020, nearly 85% of customer interactions would be replaced by machines. More financial companies are relying upon AI in greater degrees to sift through thousands of pages to spot mistakes, or atypical patterns. Despite the financial industry’s fast adoption of AI for peer-to-peer lending, stock trading, digital wallets, and robo-advisors like Watson, many financial institutions are still uncertain about security and potential data breaches.
Today, AI is actively guiding decisions across industries such as medicine, finance and manufacturing. With hundreds of layers of functions and variables tightly woven in neural networks, machine pattern and behaviour recognition is an awe-inspiring but mysterious intelligence that leaves many on the fence.
Challenges of AI in Fintech
Elon Musk, the visionary founder of Tesla and SpaceX, believes that AI is mankind’s biggest existential threat, and as such, must be held accountable and controlled before it controls—or wipes out—humanity.
To trust AI judgment or not to trust it? That is the question. The unknown is a mysterious place, creating many questions that AI innovators have thus far been unable to answer, posing the potential for it to escalate beyond human control. Fintech in particular, a sector governed by strict compliance to regulations and governance, a data breach or other security failure could be catastrophic. But despite AI’s ability to predict behaviours, disease, and trends, it cannot predict when its own failure will occur. Neither can its creators.
A growing concern surrounding AI is that if it is to mimic human capabilities, it should work according to reason and logic, and its creators should understand how that works. But nobody does. Joel Dudley, leader of an AI research group at Mt. Sinai, said, simply, “We can build these models, but we don’t know how they work.” Without understanding how AI comes to a decision or gives advice, users are essentially acting on blind faith that their machine is parsing together the right cues and data, and not becoming intelligent enough to go rogue.
As banks and financial organisations shift greater focus to the complexities of machine learning and how to incorporate it into their daily processes, questions also arise about unauthorized access to data. Financial leaders must have concrete responses for their board and executive teams when it comes to regulatory compliance within AI and Fintech.
Why take the risk?
But AI isn’t all doom and gloom. This set of technologies also has the unique ability to transform industries if we can find ways to make it’s processes understandable to creators and accountable to users. Below are just a few of the reasons many companies continue to invest in AI.
- Improved efficiency
AI positions financial institutions to streamline their heavy data processing and reporting requirements. While banks are focusing on improving efficiency and clarity with AI, many users believe that major financial decisions should still be left to humans, who have the ability to explain their reasoning and conclusions.
- Economic growth
As stated earlier in this article, research shows that by 2035, AI will have the capability of boosting economic growth by 40% in 20 countries. With increased productivity from intelligent systems, and people overseeing the technology and its applications, economies can position themselves to flourish.
- Increased accuracy
Machine deep learning, predictive behavioural analytics and data-driven marketing have the capability to eliminate guesswork from financial decisions. This security can enable commercial and private users to make more stable financial decisions.
- Risk and compliance assessment
By reviewing massive amounts of data, trades and financial histories, AI has the capability of reducing risk and increasing compliance. But this same point is the bane of many financial institutions who are wary of its security and ability to make major financial decisions without human judgment.
- Improved recommendations
An Econsultancy survey revealed that 45% of financial service executives are focusing heavily on creating personalized customer experiences. By collecting and quickly analysing data, decisions, behaviours and patterns, AI can enhance an institutions ability to provide sounder strategic solutions and advice unique to each individual.
Balancing the risks
All things considered, AI has incredible potential for reshaping and streamlining professional service firm operations and general business. But it’s evident that AI must be held accountable to ensure that it doesn’t outgrow its creators’ ability to control it. Ray Kurzweil of Google believes that AI will surpass human intelligence by 2019. To protect lifetime investments and assets, banking in general has taken a more cautious approach.
While we may have the “explainability” factor against us, we can establish boundaries in AI use by incorporating social intelligence and ethical judgments into the process. As such, AI advice and recommendations should be examined warily, in the same way that one would do for any human who can’t explain their reasoning for conclusions.
As the future progresses, AI will be applied more and more to creating new models and processes, dealing with rules and exceptions, reporting earnings, “closing” and reconciling data and accounts, forecasting earnings and allocating assets. To implement AI wisely, leaders must realize that it should not be applied to every aspect of their business. Financial firms should make AI an extension of their existing analytics team, to strike a balance between human and machine. To eliminate data breaches of an intelligent system that can’t be explained and perhaps fully contained, they must implement secure, agile systems that reside outside of machine learning systems. Centralizing data in external server solutions can enhance financial security and peace of mind.