Regulators Should Keep Pace with the Use of Artificial Intelligence in Financial Services, Report Finds

Artificial Intelligence ("AI") has the potential to bring about dramatic changes in the financial services industry, a new survey has found. 

Such developments could help financial institutions improve their compliance and screening programmes and remove the potential for human error, although the adoption of these solutions will hinge on regulators' ability to keep pace with technological changes. In the Asia-Pacific region, the majority of survey respondents expressed doubts about whether local regulators were up to the task. 

The survey, conducted by law firm Baker & McKenzie, found that 75 percent of the companies questioned did not believe regulators had kept pace with the immense technological advances made during the past decade. The survey found that, a result, firms in the Asia-Pacific region had been cautious in their own adoption of such technology. 

Respondents called for closer collaboration between regulators in the region and fintech developers. 

"This is a call to action for all of those with a stake in fintech, including start-ups, financial institutions, regulators and consultants such as those in the legal profession," said Gavin Raftery, a partner with Baker & McKenzie in Tokyo. "At a minimum, new technologies need to be understood as they come on stream, and information sharing and collaboration needs to be embraced to keep pace." 

The report said that AI could bring about competitive improvements in the financial sector, as it presented the opportunity for software to learn from patterns and behaviours and act accordingly. 

As one of the fastest growing areas in fintech, such technology offered competitive and innovative advantages to early adopters, the report said. Firms in the regional fintech hubs – Hong Kong, Singapore, Sydney and Tokyo – might therefore be better placed to benefit. 

Raftery said new market entrants would be able to disrupt the financial services world by harnessing AI for platforms as diverse as investment research and risk management. "That is why we see many financial institutions actively involved in incubator and accelerator programmes: to stay abreast of change, identify investment opportunities and to tap into top talent," he said.

Compliance Uses

For the compliance, legal and risk management fields, AI was an opportunity to automate various functions and reduce conduct risk, Raftery said. 

"This is a very topical issue for many financial institutions as there have been some very large fines handed down globally for misconduct and compliances breaches by traders," he said. 

Indeed, such technologies may already have helped the compliance function to deal more effectively with anti-money laundering, know-your-customer and countering terrorist financing issues.

"AI has tremendous uses in the AML, KYC and CTF space, especially for customer onboarding and ongoing transaction monitoring, because it can pick up on patterns and trends in real time that even an analyst with years of financial crime experience may not detect in time, if at all," said Bill Majcher, chief executive of EMIDR in Hong Kong. "At the end of the day, [analysts] are only human … Computers are not perfect, but solving for overheating and power surges, they can keep grinding on, searching for patterns while the rest of the world sleeps." 

Raftery said, however, that AI would need to be introduced carefully, and that thorough testing was needed to ensure that institutions, regulators and policymakers alike understood the potential downside of such technology. "If and when introduced, the role of compliance officers and in-house counsel will also need to evolve and the integration of tech-savvy personnel into these teams will be important," Raftery said.

As an example, high-frequency trading ("HFT") is predicated on algorithms which employ AI, but developments in AI have taken this technology beyond simply enabling fast transaction execution.

"Many fintech start-ups are now working with AI on various forms of machine learning, for example to detect trading and market sentiments. It is important for those developing AI platforms to work closely with regulators to ensure transparency and the sharing of knowledge and experiences," Raftery said. "Regulators will not easily authorise systems and will come down hard on market participants that introduce new risks to the market."

Looking Ahead

The survey polled 424 finance executives, of which 67 came from across the Asia-Pacific region. Those polled saw AI mostly as an opportunity for financial services providers, although one not without inherent risks. 

Survey respondents said that over the next three years, the impact of AI would be felt most keenly in areas such as trading, financial analysis and information technology. Many participants also expected machine learning to affect risk assessment, credit assessment and investment portfolio management. 

Risk assessment and financial research were the areas where firms said they were most likely to invest and experiment with machine learning applications in the near future. Half of the international financial services companies interviewed expected some level of AI to be introduced into their risk assessment functions during the next three years, although in Asia that figure was just 38 percent. The cost of development remained the main obstacle for building AI or machine learning capabilities, followed by a shortage of skills to develop and maintain AI systems.

Survey respondents also said AI would encourage market diversity, with more start-ups and small and medium enterprises ("SME") entering the market. Some, however, expected this to have a negative effect on market stability as new entrants disrupted the established order. 

More than two-thirds of those polled said their own roles in financial services would be substantially or completely changed by AI and machine learning within the next 15 years.

New Technology Cannot Completely Remove the Potential for Misconduct 

Some aspects of the financial services world, however, may be more resistant to change. The report found that despite the introduction of new technology, it would be difficult to remove completely the potential for misconduct by traders. 

"Financial institutions have been fined billions of dollars because of illegality and compliance breaches by traders. A logical response by banks is to automate as much decision-making as possible, hence the number of banks enthusiastically embracing AI and automation," said Arun Srivastava, a partner with Baker & McKenzie in London. "But while conduct risk may be reduced, the unknown risks inherent in aspects of AI have not been eliminated."


Ajay Shamdasani is a senior staff writer with Thomson Reuters Regulatory Intelligence in Hong Kong. He covers regulatory developments in Hong Kong, India and South Korea. He also writes about money laundering, fraud, corruption, data privacy and cybercrime.