While the reskilling revolution is sweeping across the financial ecosystem, there is also a growing need for digital ethics that will ensure that major financial institutions efficiently bridge the technology gap and enable staff to enhance their new careers.

As we embrace convenient and flexible digital services, inside and outside of financial services, the boundaries of what we share and who we trust are shifting. With regulatory initiatives such as GDPR enabling individuals to exert greater control, banks must balance customers’ needs and rights with their duty and obligation to protect their assets, including data.

Ahead of SIBOS in London, Finextra spoke to leading figures in financial services Craig Macdonald, digital transformation lead at Accenture; Androniki Menelaou, innovation lead at ING; Salla Franzén, chief data analyst at SEB; Clare Gambardella, Chief Customer Officer, Zopa, Prema Varadhan, chief architect and head of AI at Temenos and Karan Jain, fintech & regtech advisor about AI and the role of the human in financial services.

Beyond augmenting compliance

Where are we at with AI today? Have financial institutions started to use it for one its most proven use case: augmenting compliance? How far away are we from an AI regulation being put in place?

Macdonald highlights that the financial services sector has now entered the “intelligence age” but is “doing so under intense pressure on multiple fronts. Rapid advances in AI are coming at a time of widespread technological and digital disruption and competition is fierce.

“AI is taking this disruption to a new level, with 43% of global banks adopting it across a wide variety of use cases, ranging from credit decisioning to customer service AI powered virtual agents. With AI-augmented operations, a bank can expect cost savings of between 20 and 25%,” Macdonald refers to statistics from Accenture Technology Vision 2019.

However, as Macdonald describes, AI raises concerns about how disruptive the technology will be. “These fears include workforce displacement, loss of privacy, potential biases in decision-making and lack of control over automated systems and robots.”

He continues to say that banks now have the responsibility of helping their workforces understand how AI can help enhance productivity and job satisfaction, which will reassure them that they have a crucial role in the future business, working alongside the technology, rather than having been replaced by it.

“Part of this will indeed include a framework for regulation and compliance, such as building auditable and regulatory-compliant AI solutions, and identifying illicit behaviour more accurately and efficiently.” Before focusing on regulation and compliance closely, it is important to reiterate in addition to reassuring the workforce, AI can play a major part in the analysis of data and how this impacts the core of the business.

Franzén reveals that “using data for different needs and purposes becomes a collaborative exercise, where data scientists and business professionals can join hands with IT engineers to create the future of banking. Compliance, risk, business development, daily business - the data is a common asset and should be shared in a secure and integrity-preserving way.

“In compliance the opportunities to augment the current practices are vast. Moving from using simple expert systems to more complex ones, such as dynamic graph and network analysis, natural language processing and behavioural modelling, is quite a journey and requires retraining of employees to equip them with the tools of the future.”

Today, AI, automation and machine learning are being used in the financial crime compliance space and is changing the way that compliance professionals work. Franzén adds that AI regulations are being discussed in many different fora, with the High Level Expert Group (HLEG) in the EU releasing guidelines and suggestions for regulatory actions within the European Commission.

“Balancing ethical considerations, integrity and valuable services will be one of the main concerns when implementing sustainable AI in financial services. Not only do bankers and clients need new knowledge, but also legislators, citizens, and society as a whole need to understand the risks and possibilities with these new tools. Parts of AI will need to be regulated, and it remains to be seen how granular the legislation becomes.”

Jain has a similar attitude and claims that “AI or machine learning has swept quickly through the compliance space since 2016 as mainstream. Compliance is in large parts, can be a time-intensive and repetitive process, and has traditionally been very manual process as the industry gets to grip with what does good compliance look like. The implementation of regtechs that digitise these processes is the first step to realising augmented compliance.”

He continues to say that regulators such as the UK’s Financial Conduct Authority (FCA) and US counterparts are also getting in on the action with initiatives on digital regulatory reporting such as the FCA Tech Sprints, that are investigating what digitised regulation would look like and promote deep tech to fight financial crime at scale with AI and data sharing.

“Equally, the power of AI developed by some technology companies is beyond comprehension. If unregulated, it has the potential to causes major social unrest in the future, essentially creating a super-class enabled by AI and technology.

“It is a sensitive matter, and regulators have been in active discussions with the private sector to ensure balanced approach is applied which doesn’t impact innovation at the same time. I can't say for sure when an AI regulatory framework will be put in place, but I would say in the next 12-24 months we should see something in place which will improve over period of time.” Perhaps in time for SIBOS 2020.

Robots do not have that ‘je ne sais quoi’

If robots do not have a particular ‘je ne sais quoi’, how much will a human’s role in banking truly change? Will kind of new roles will be created? Will financial institutions have to change their modus operandi to keep up with digital transformation? Will the reskilling of humans mean that there will be a diverse range of people employed at banks?

It is evident that following efficient retraining, humans have a crucial role in the future of financial services and will be working alongside the technology, rather than having been replaced by it. In turn, employees will be elevated and given more rewarding and higher-value roles, resulting in a true redefinition of the sector as we know it today.

As Franzén does, if we were to compare the world of 20 years ago to where we are now, life without the internet is difficult to imagine, “yet the world wide web is not even thirty years old, and household Wi-Fi has only existed for twenty years.

“The ever-changing technology landscape means that banks need to increase their speed of adoption of new technologies while taking care of their legacy footprint. New roles, such as data scientists, social media communicators, UX programmers are only the beginning of the transition to something new.”

In addition to this, Jain points out in order to reap the greatest rewards from the combination of humans and emerging technology, “we must consider the strengths of humans and computers, allocate tasks accordingly. Despite initial resistance, when the calculator was invented, humans quickly accepted benefits and adopted to other value add activities.

He continues: “We are soon going to forfeit most of our forecasting and prediction ability to machine learning. The studies consistently show that humans significantly overestimate, this is a form of hubris in a way, and we will find much greater productivity when we accept that AI is much better suited to these types of computationally intensive tasks.”

Again, while humans will not be replace, in order to be effective in the future, financial services employees must focus on becoming technology-enabled. Jain also says that people are good at “abstract, experiential judgement and creative tasks that require a reasonable level of dynamic, lateral thinking.” Robots do not have this ‘je ne sais quoi’.

However, Menelaou states that “we need to remember that robots are not self-created, the robots will be trained on ‘human behaviour’ or the ‘optimal behaviour’ and will try to replicate it.” Humans are good at being creative, robots are better at deriving a decision from hundreds of points. Although robots are unable to identify the right problem to solve, what type of products to build, how to build effective teams and how to innovate, according to Menelaou.

Varadhan adds that a key area here will be automation, where manual tasks that were once done by a human can now be conducted by a machine, which can automate those processes, freeing up the human to focus on responsibilities that require a personal touch.

“With increased automation and increased efficiency comes increased bandwidth and an evolution of what it looks like to be and act as a bank. There will be more skills required to maintain automation and create clever use cases instead of conducting traditional software development.

“As AI automates tasks and conducts tasks, employees can be upskilled and re-employed. This is only really required in niche areas though - we will see more data scientists and automation engineers, but the banking roles specifically won’t necessarily change,” Varadhan says.

The emergence of behavioural banks

Will the evolution of the bank progress from neobank and challenger bank to behavioural banks? What are behavioural banks and how soon before we start seeing them on our high street, or app stores?

To riff on Varadhan’s point, automation will lead to a remodeling of the workforce and a transformation of what it means to be a bank, resulting in the emergence of behavioural banks. But what are behavioural banks?

Macdonald highlights that: Behavioural banks are firms which introduce dynamic interest rates that are linked to financial behaviours of their customers. Similarly to the use of Fitbits in personalised insurance, this enables those customers to earn more interest on savings and pay less interest on credit as they improve their financial behaviour or monitoring how people drive to adjust their insurance.”

To use Zopa as an example, Gambardella explains how the financial services company uses customer data, combined with credit bureau data, to calculate individual loan rates and aim to leverage this system in the credit card and savings market.

“Currently, behavioural insight is limited by the data available. With greater adoption of open banking, customers will be able to allow financial providers to see a holistic view of their financial data, which in turn will lead to the concept of true behavioural banking becoming more viable.”

Menelaou has a similar view and says that “banks need to win this relationship with their customers by providing products and services that engage with customers and respond to their needs, including products outside the realm of banking as we know it.