The regtech revolution is upon us

Numerous theories have emerged as to why the UK has endured such a severe productivity problem over the last decade. Growth in productivity of just 2 per cent during that period would typically have been accomplished within a single year prior to the financial crisis, and thus the 'productivity puzzle' continues to confound economists. In all likelihood, there are a number of reasons why productivity rates have stagnated, but one of the most compelling has been the increase in regulation, and the resultant compliance burden on companies. 

Clearly, after the so-called 'casino banking' which precipitated the global meltdown, such intervention was a necessity. In the UK, the Financial Services Authority made way for the revered Financial Conduct Authority (FCA) and the Prudential Regulation Authority (PRA) in 2013, and most experts agree that the banking system is now far more stable, and better capitalised to deal with any future economic headwinds.

Yet, as regulators continue to flex their muscles, an ever-bigger army of legal, compliance and risk assurance specialists are needed within companies. Until now, it's been a labour-intensive line of work - both in terms of the watchdog enforcing the rules, and the professionals adhering to them. The latter play a vital role, but, from a productivity perspective, there is little in the way of value created in the economy by such workers, and the expansion of this niche serves to further undermine productivity rates.

Poor productivity may seem like a reasonable price to pay for a safer financial system. But what if the two weren't mutually exclusive? Or at least, not to the same extent that they are now? 

When machines and regulators meet

The UK has been a pioneer with integration of artificial intelligence (AI) and machine learning (ML) into regulatory technology - or 'regtech', as it is commonly known. Last year, the FCA, in conjunction with the Bank of England, launched its Digital Regulatory Reporting (DRR) initiative – one geared towards developing machine-readable reporting, and using technologies such as Recurrent Neural Network and semantic web to forge simple, more-efficient processes for financial institutions to comply with regulatory requirements. Specifically, it sought to establish reporting rules which are more automation-friendly, according to BankingTech.

Eliminating the need for human interpretation for such reporting has the double benefit of both cutting down the cost burden for companies, and also the number of reports they are obliged to submit. Much of the FCA's focus was on how best to dovetail this machine-readable regulatory framework with a standardised language. And this venture has already yielded success, with the FCA confirming late last year that it had integrated machine-reading technologies into two of its regulations (related to mortgage lending criteria and capital requirements), and expansion into other regulations has already begun in 2019.

The role of AI, semantics and humans

This may only be the very genesis of the regtech revolution so far, but it represents impressive progress. That said, the exponential gains will undoubtedly come through AI and ML. The rate at which these technologies can process and store data is a game-changer, and advanced algorithms can even provide outputs and analysis which resemble human thought. As such, AI has the potential to go beyond the menial, and provide real-world, problem-solving capabilities.

Semantic web, which involves a network of web pages developed and tagged in a standardised way that is readable by machines, could well be the main catalyst for the rise of AI within regtech. The one barrier to semantic web is the human dependency in developing the structured data which underpins this language, and the difficulties this presents when it comes to scaling up. However, the rise of cognitive computers may yet help to automate the development of this structured data through AI-based algorithms, and pave the way for semantic technologies to power the regtech revolution.

Whichever medium of AI comes to the fore, the role of human expertise most certainly won't become obsolete in the near future. Ramping up efficiency, and reducing the need for professionals to carry out mundane tasks is the basis for a successful marriage between AI and regulation. Yet optimising the integration of AI and ML, and maximising their capabilities, will perpetually require a human touch.

And of course, in a fast-changing world, ongoing oversight to identify risks, and ensure existing rules are fit for purpose will always be beyond the scope of robots. The very concept of compliance itself is also contingent upon the compass of morality, and this should not be forgotten either.

Keeping up with the Zuckerbergs

Nevertheless, in a world of disruptive technologies and borderless financial services, the rapid evolution of regtech is a pre-requisite if regulators are to keep pace. And that goes for companies too. In the UK alone, firms must comply with the standards set out by the likes of the FCA, the PRA, the Payment Systems Regulator, the Competition and Markets Authority and the Open Banking Implementation Entity. Many such rules are not aligned either, so the burden is a considerable one. Factor in firms who operate at a global level; dealing with a whole new set of regulators in other countries, and the scale of complexity becomes all-consuming.

This underscores the importance of integrating AI into compliance in a coherent way which is adopted uniformly by watchdogs and central banks the world over. There are considerable challenges on the way - not least from the likes of Facebook's new Libra crytocurrency, and the myriad of loopholes regulators will need to plug within this particular niche. It is only through artificial intelligence and machine learning that regulators will be able to get ahead of the curve with such trends - whilst still promoting efficiency, and maintaining an environment where compliant, law-abiding firms can prosper.