Will the evolving use of artificial intelligence in financial services hollow out the sector by replacing human beings with automated functions and algorithms? If not, what role will humans play in an increasingly AI-driven future? Tariq Munir, a Digital transformation advisor and keynote speaker, and Gilles Bonelli, founder and CEO of UK digital and sustainable finance consultancy See The Next Move, discuss with VitalBriefing thhe disappearance of jobs in manual and high-volume tasks such as loan approval offset by a refocus on roles in which human interaction is essential.
In which areas can AI most obviously make a rapid improvement to the efficiency or cost of financial industry functions?
Tariq Munir: AI is already creating efficiency and reducing costs in several areas within the financial industry, for instance fraud detection. Machine learning algorithms are being used worldwide to detect credit card-related fraudulent transactions in real time. Similarly, malicious activities by scammers can be flagged in real time through better detection algorithms, saving potential victims billions of dollars.
Another area is algorithmic trading – by analysing patterns in vast volumes of market data, algorithms can make better investment trading decisions. It can boost the efficiency of operations and enable financial institutions to make better decisions, optimising customer journeys and internal operations. For instance, Singapore’s DBS Bank has used AI to create predictive models that reduce ATM downtime and optimise cash replenishment, decreasing the incidence of machines being out of cash instances and the frequency of restocking.
Gilles Bonelli: One key use case is AI chatbots to address customer inquiries in retail banking. Chatbots can handle basic customer inquiries around the clock, freeing up human agents for more complex issues and reducing operational costs. Another is real-time fraud detection systems in payment processing. AI systems can detect and block fraudulent transactions in real time, enhancing security and reducing financial losses.
The insurance industry can draw on predictive analytics for risk assessment, using AI to analyse historical data and current trends to predict potential risks, making for more accurate pricing and better risk management strategies.
What areas of the industry does AI appear most likely to reshape in the longer-term period from 2025 onward?
Tariq Munir: I see the convergence of technologies. AI can play a big role in the evolution of decentralised finance to create an autonomous, sophisticated and universal financial system.
As both the volume of data and our capacity to compute increases, advanced AI models will become better at macroeconomic prediction, revolutionising monetary policies and investment strategies, and possibly empowering regulators with better insights to draw up industry rules. We might just be able to predict financial crises with relative certainty and avoid them thorough pre-emptive measures.
Similarly, the convergence of the metaverse with the mainstream financial system would mean smarter algorithms conducting banking functions in immersive and virtual environments, connected to real-world financial systems. These complex networks can only be managed using powerful AI systems, revolutionising how users interact with banking networks.
Gilles Bonelli: In the longer term, AI-powered robo-advisors in wealth management can provide automated, personalised investment advice, making wealth services more accessible and efficient. It can also transform credit scoring in consumer lending, incorporating non-traditional data sources such as social media activity and utility payments to provide a more comprehensive assessment of creditworthiness.
It can also transform regulatory compliance through use cases such as automated compliance monitoring in investment banking. AI will be able to monitor transactions and communications continuously for compliance with regulations, reducing the risk of breaches and penalties.
To what extent might AI eliminate tasks currently carried out by human beings, or could it be a tool to help them carry out tasks more effectively?
Tariq Munir: The future of AI and humanity is one of co-existence as opposed to replacement. It will automate a number of tasks, but in essence, it is still an augmentation rather than a replacement for human intelligence.
Effectively, AI automates the bulk of manual and high-volume tasks (as in loan approvals, for instance) and pushes humans towards roles that require human interaction. A multi-agent model can resolve the majority of customer complaints, creating massive efficiencies for humans to focus on strategising and building meaningful customer relationships.
Gilles Bonelli: The tasks AI can eliminate include automating data entry in accounting, eliminating manual input and reducing human error. In corporate finance, AI tools can provide analysts with deep insights and predictive analytics, enhancing their ability to make informed financial decisions.
What role can AI play in the democratisation and personalisation of financial services offerings, and improving institutions’ decision-making?
Tariq Munir: AI-based solutions are also enhancing customer experience, including use of sentiment analysis and real-time analytics to understand preferences, increasing satisfaction and reducing customer churn.
An example is the loan approval process of Ant Financial’s MyBank in China, which uses data from Alipay and other Alibaba group services to approve loans for SMEs using so-called 3-1-0 digital lending model – completion of loan applications by SMEs in the 3 minutes, obtaining approval in one second with zero human touch.
Gilles Bonelli: An example of the potential for personalisation is customised loan offers in consumer banking. AI can analyse individual customer profiles to offer personalised loan products, improving customer satisfaction and conversion rates. Another is analysing market data in real time to support trading decisions in investment banking, increasing the agility and profitability of trading operations.
What are the overall implications of AI for financial sector employment? What kind of jobs are most likely to disappear, and in what areas might employment increase as a result of use of AI?
Tariq Munir: It will give rise to entirely new career paths for humans, for people who can interact with algorithms both at the surface – such as prompting Generative AI models – and at the back end, in building and researching new models.
In addition, roles that require business partnering with other functions and involve human interaction will become more valuable. In a world where algorithms run operations, the value of human skills such as empathy, critical thinking, emotional intelligence and leadership will only increase. If anything, AI will make us more human.
Gilles Bonelli: Among the developments that will see systems replace humans is the automation of routine transaction processing in retail banking, which will lead to the disappearance of clerical roles. By contrast, the development and maintenance of AI systems will create new job opportunities in software development, data science, and cybersecurity.
In what areas of the financial industry might the tendency of generative AI models to ‘hallucinate’ carry particular dangers?
Tariq Munir: Any area in which generative AI is infused can pose the risk of plausible-sounding but incorrect outputs by these models. Areas including credit scoring and market analysis, risk mitigation and insurance claim handling, can all be impacted by models hallucinating.
However, it is important to understand the difference between hallucination as opposed to generative AI models providing incorrect output simply as a result of bad prompting. An incorrect output is not necessarily a hallucination, although every hallucination is an incorrect output. Building AI and digital literacy is the only way to recognise and counteract these effects.
Gilles Bonelli: Risk areas include AI-driven decisions on loan approvals in consumer lending, where incorrect AI-driven credit assessments could lead to inappropriate loan approvals or rejections, impacting financial stability and undermining customer trust.
Another is the use of trading algorithms in investment management. Where hedge funds employ automated trading systems, erroneous signals from AI systems could cause significant financial losses if trades were executed based on hallucinated data.
Could AI help to reduce the risk of malicious hacking of financial institutions or clients, or might it offer new tools to fraudsters and other wrongdoers and create new vulnerabilities they might exploit?
Tariq Munir: It is a tricky situation – AI offers the potential to overcome the risks of attacks by bad actors, but also provides immense power to exploit vulnerabilities. AI containment is a complex issue that requires a concerted effort between researchers, companies, governments and the wider community to build guardrails around what can and cannot be developed, which models cannot be released as open source, and what fail-safe measures we need to build to ensure decisions made by AI on our behalf can be reversed.
The biggest issue with containment is access to AI models. With conventional weapons of mass destruction, a wrongdoer needs access to multiple technologies and infrastructure to create something that can endanger humanity at a large scale. For example, not everyone can build a nuclear plant. But with AI, anyone with access to sufficient computing power and harmful intent can jeopardise a lot of people and systems with relative ease.
Gilles Bonelli: In digital banking, AI can identify and respond to cyber-threats more swiftly, enhancing the security of financial institutions and their customers. But while it can prevent fraud in payment processing, it can also be exploited by fraudsters to identify new vulnerabilities, requiring continuous adaptation and improvement of security measures.
Could heightened public perception of the automated nature of financial industry processes spark fresh resentment of the industry and risk a populist backlash?
Tariq Munir: With every change, there is always apprehension and backlash. However, I reiterate that the current narrative of AI replacing humans is inherently incorrect and misleading. If anything, it is enabling processes to be faster and better. I believe AI will lead to better trust in the financial system by providing a democratised and possibly decentralised way we interact with financial institutions.
Gilles Bonelli: Potential exists for resentment in areas such as automated customer service systems in retail banking. This might lead to customer dissatisfaction if it is perceived as impersonal or inadequate. However, implementing transparent and ethical AI practices in financial services, along with clear communication about the technology’s benefits and limitations, can help mitigate negative public perception and foster trust.
— — —
Tariq Munir is a Digital Transformation advisor, an international keynote speaker and LinkedIn Top Voice with 20+ years of experience spanning Finance, Integrated Planning, and Digital/Finance Transformation. He has worked with various Fortune 500 companies and brings expertise in leading complex and large-scale Digital and Process Transformations. He can be reached at hello@tariqmunir.me, on his through (www.tariqmunir.me) or via his LinkedIn profile.
Gilles Bonelli is a seasoned Management Consulting Executive and qualified Accountant (FCCA) with a distinguished career in Digital Transformation, specialising in Finance, FP&A, Shared Services/GBS, and Centres of Excellence. He is recognised as a Top Voice in AI on LinkedIn and an Influential CEO in Data Science (CEO Monthly). Gilles has a deep expertise in integrating AI into financial processes, leveraging cloud-based ERP/EPM/BI systems, and reshaping business operating models. He can be reached at gilles.bonelli@seethenextmove.com or via his Linkedin profile.
