The Human Factor And AI In The Finance Sector

How can Extract AI and Automated Machine Learning be used in the finance industry.

The human factor and artificial intelligence in the financial services

Telesurgeon. Garbage designer. Augmented reality tour provider.

Conjuring new job titles which don’t currently exist can be a fun exercise in stretching the boundary of your imagination. The task becomes less cheery, however, when one realises that by implication, an equal number of jobs currently in existence will be rendered redundant by automation and algorithms. The incessant wheel of technology works like a wave — bringing in new innovations by necessity disrupts and displaces. Things become interesting when we examine particular fields through this lens. The financial services, for example.

Ah, that shining beacon of hope to legions of fresh grads hoping to make a killing, on Wall Street or its global equivalents. Nonetheless, the requirements imposed by said industry upon these grads, the skills that are valued and looked for, are all changing faster than they ever have. Artificial intelligence — along with a slew of emerging technologies like blockchain, analytics and internet of things (IoT) — are revolutionising the financial services as a field. Banking, insurance and accountancy will likely be rendered completely unrecognisable within the next decade by the impact of AI in areas as varied as underwriting and fraud detection or even tasks as mundane as ATM maintenance.

How, you ask?

In a world where the cloud becomes a vast storage facility for information, a technique called natural language processing has become incredibly powerful. AI algorithms sift can now sift through billions of words on social media and document archives for a myriad of purposes. Sentiment analysis for the trading floor, credit-scoring customers based on what they post, training customer service chatbots. Machine vision can digitalise archives at a rate comparable to a whole building of copywriters and Xeroxes. The best fraud detection and anti money-laundering measures all employ artificial intelligence.

The list goes on.

For listicle writers, this is the stuff of legends. An inexhaustible repository of examples of how artificial intelligence is completely changing the financial services because change is everywhere you look, from the front desk at your local bank to the trading floor at Goldman. Forbes and Emerj both provide a more comprehensive understanding of how this change manifests itself in the financial services. These articles offer largely conceptual explanations. While it’s great to have a more numerical grasp of the relationship between AI and the financial services, it is all too easy to lose track of the human factor underpinning this surge of change.

In fact, if we look beyond the immense jump in capability artificial intelligence can offer in countless use cases, the whole process by which the financial services will be transformed could quite possibly be shaped by a few predominantly human factors.

The thirst for talent

Because “behind every successful algorithm is a human.” A modern twist on the slightly sexist saying of old which just brings any discussion of change back to jobs and their nature. A fundamental problem with integrating AI into any business is the cost and scarcity of the relevant expertise. The demand for data scientists and AI experts far outstrips supply; a PwC study earlier this year showed that almost a third of financial executives were worried about their inability to meet the demand for AI skills in their organisations. These are large multinational firms with the resources at their disposal to invest heavily in AI. Outside of this exclusive club, accessibility itself becomes a problem. Small and medium players in the financial services are less able to leverage on AI, if at all, because they cannot afford the expertise.

The bottom line is that humans are still the key to integrating AI into any business, and right now there are neither cheap nor abundant.

Startups are trying to tackle this accessibility problem, especially in emerging markets. Extract AI, for example, uses a species of algorithms known as automated machine learning for businesses to harness the predictive capabilities of AI without directly (and expensively) employing AI experts.

Beyond cost: A new differentiator?

One of the key findings from the World Economic Forum’s latest report regarding the implications of AI on the financial services industry involved the introduction of new differentiators besides cost. In watered-down terms, this means that customers might become less price-sensitive as AI diversifies the range of customer experience. From automated chatbots to AI models that help customers make financial decisions, industry players like banks and insurance providers will become more motivated to distinguish themselves to customers instead of engaging in “race to the bottom” price wars. The use of AI in the financial services will be human-centric and not data-centric, with customer satisfaction being the ultimate objective.

Explainable AI (xAI)

As the decisions made by AI algorithms become increasingly complex, the risk exists that humans will be unable to explain the decisions made by AI models. Explainable AI is a set of methodologies to ensure that any such decisions can be easily trusted and understood by humans. One such technique is the use of counterfactual predictive models where when a decision is made, a sort of quasi-twin input is generated with the closest possible parameters but a different decision.

For example, when a consumer loan application is denied, the counterfactual model generates an application resulting in approval which is as similar as possible to the original applications—say with a marginally higher disposable income. By inference, the differences in the counterfactual model were the most sensitive factors in the rejection, thus explaining, at least in part, the decision made by the AI model. In our highly oversimplified example, the disposable income was too low for the loan application.

The human factor

Human expertise, customer service and explainability are all vital elements of AI’s role in the financial services. Artificial intelligence might be reducing human input in the provision of financial services, but the human factor will still very much determine the manner in which AI shapes the industry.


Extract AI

Extract AI aims to democratise the AI industry, allowing any business to leverage on the benefits of AI. Together, we are building a world where you can predict tomorrow.