Ghost in the Machine: Financial Regulators Keep an Eye on AI

The increasing applications of artificial intelligence and machine learning in financial services – from robo advisory platforms to algorithmic stock trading to risk modelling – is disrupting decades old processes. As the industry embraces AI-led innovation, can regulators keep up? And more importantly, how might technology solve some of their biggest challenges?

Across most industries, adoption of artificial intelligence (AI) shows no sign of slowing. Financial services is no exception: asset managers, banks, and other financial institutions are using AI and machine learning to streamline operations, manage assets, and bolster cyber security efforts. But as the use of AI in the industry increases, it raises the question whether regulations are sufficiently fit for purpose when the activity they are regulating involves both human decision making and AI. Several recent developments suggest regulators are proactively assessing these technological advancements both in how to appropriately regulate them and how to deploy technology as a resource.

“I think, therefore I am”

Just last week, the UK’s Financial Conduct Authority (FCA) said they could use AI and machine learning to detect non-compliant advisers. The FCA said they are exploring investments that would use publicly available data, such as complaints and past records, to identify advisers who could be likely to mis-sell. In 2016, the FCA set up a regulatory sandbox, approving 24 projects that focused on robotics and distributed ledger technology (DLT). Earlier this year, the FCA announced plans to introduce machine-readable regulations. Nick Cook, the head of FCA’s data and information, stated they intend to “put out rules which are written manually in ways that can be fully and unambiguously interpreted by machines.”

The US Securities and Exchange Commission (SEC) is also making strides to embrace new technology and to provide an open forum for industry participants. Earlier this month, it launched FinHub – a centralized resource to engage with the public on fintech related initiatives including automated or “robo” investment advice and AI. The SEC said they aimed to provide a portal for the industry and the public to “engage directly with SEC staff on innovative ideas and technological developments.”

Bob, meet Bot

Regulators are also using AI to create efficiencies of their own. Funds, products, and structures are more complex than ever before, which in turn increases the burden of regulating those entities. AI can help regulators streamline their jobs and sort through massive quantities of data – essentially do more, and more accurately, with less.

Consider credit card companies: many use AI to sort through millions of transactions to flag potential suspicious activity. Likewise, regulators can use AI to differentiate between firms that are likely to generate notable breaches and those less likely to do so. For example, the FCA is using AI to generate a “risk score” for the entities it regulates and can then sort firms into different risk buckets. Part of the difficulty of regulating this technology is that an AI system can only show results but not explain how it arrived at the conclusion – often without creating an audit trail. Regulations are predicated by discharge of fiduciary duty, discretion, and conscious decision making, and as such, use of AI in certain areas of a firm would not satisfy regulatory requirements as they stand now.

Additionally, while a human can choose to ignore an irregularity in the data and apply reasoning and rationality to an anomaly, a computer cannot. When a computer makes a decision on behalf of a manager, what standard should regulators use to scrutinize that decision? For now, regulators believe individual human accountability remains and use of any AI forms part of a wider risk management framework. So, the entities they regulate cannot outsource or blame a robot for any failings. But as AI evolves, this thinking may too take new shape.

Adapting to new technology is the new normal and this is only the beginning. It’s important for regulators and the industry to keep an open dialogue about what technology they are using, how they are using it, the problems technology is solving, and the problems technology is creating. The OnTheRegs team is off to grab coffee with Siri, Cortana, and Alexa, but we’ll have much more to say about AI and other fintech-related topics coming soon.