Although artificial intelligence (AI) is set to improve financial institutions’ risk management, credit provision, asset management, trading and hedge funds, certain risks will remain, says multinational law firm Baker & McKenzie banking and finance partner Chris Hogan.

These risks include potential malfunctioning algorithms, human misuse of technology and concerns surrounding the security, privacy and quality of data.

Some companies are pioneering AI trading programs for trading and investment management, which use a combination of machine learning techniques and evolutionary algorithms to process massive volumes of data and recognise obscure patterns.

Baker & McKenzie’s survey, titled ‘Ghosts in the machine: Artificial intelligence, risks and regulation in financial markets’, found that many of these AI software programs learn and update their models automatically and independently from human interference.

Machine learning is also predicted to make financial markets more competitive as soon as 2018. However, financial services companies are increasingly concerned about the risks and ability of regulators to keep pace.

“The vastly increased algorithmic powers emerging can help to build economies by closing procedural gaps, and create opportunities,” says Baker & McKenzie intellectual property and disputes partner Darryl Bernstein.

However, much of the financial technology is still in the experimental stages in many applications, including trading, portfolio management and credit assessment. Therefore, the possible risk of malfunctioning algorithms, together with concerns surrounding the security, privacy and quality of data, has led to calls for new regulation, he emphasises.

Survey respondents also expressed unease about the future regulatory response to AI. More than 75% are not confident that regulators have adequate knowledge and skills to stay abreast of new financial technologies and understand the potential implications of AI for financial markets.

Survey participants believe that regulators are only just beginning to understand the potential implications of AI for financial markets and companies. The report notes that any rule writing on machine learning in the next few years is expected to focus on the integrity of algorithms, on which attention is only now being focused.

Meanwhile, the rise of AI in the financial services sector will also lead to job losses in certain areas.

“Local banks are already reducing employment, as they become more digitally enabled. Computing technology enhances data processing significantly. The introduction of different business models will lead to some jobs disappearing and others developing,” says Bernstein.

In turn, there is a need to develop education systems to embrace this, as is the case in Asian markets, with the government of Singapore acting as a bridge between academia and business in the financial technology space. South African authorities must play an enabling role for technological development in the local market, and fulfil the role of regulator.

“However, based on the delayed enactment of the Protection of Personal Information Act, we do not expect fast, widespread financial technology regulation in South Africa,” Bernstein concludes.