While Gen AI holds big promise for banking, most of the current deployments are limited to just a few areas or don’t go beyond the experimental phase. Though early pilots appear impressive, it will definitely take time to realize Gen AI’s full potential for the banking industry.
Leaders in the banking sector must address significant challenges as they consider large-scale deployments. These include managing data security, integrating legacy technology, navigating ethical issues, addressing skills gaps, and balancing benefits with regulatory risks.
Five Challenges for AI in the Banking Sector
1. Data Security Concerns
The adoption of Gen AI raises data privacy and security concerns, which are major issues for the banking sector.
First, there is a risk of unintentional violation of privacy rights when collecting large amounts of client data for profiling and forecasting, even if the data is already publicly available. Gen AI could inadvertently reveal sensitive or personally identifiable information, such as personal identification details, transaction history, and account balances.
Gen AI also provides a new tool that fraudsters could use increase the sophistication and scale of their scams. For example, scammers could use Gen AI to quickly create new phishing attacks, smishing attacks, fake browser extensions, or impersonation scams.
To mitigate data security risks banks should deploy robust cybersecurity measures to prevent hacking attempts and data breaches.
2. Legacy Systems
Legacy technology is another factor slowing down Gen AI’s commercial use. Such systems impede the adoption of novel technologies and the integration of the new capabilities that these innovations can deliver for several reasons. First, legacy systems often use outdated data formats, structures, and protocols that may be incompatible with modern AI technologies. Secondly, they may store data in siloed or proprietary formats, making it difficult to access and retrieve data for AI model training and analysis.
Interestingly, Gen AI itself can serve as a solution to the legacy infrastructure problem by accelerating the transition from legacy software and data storage, which previously seemed cost prohibitive. Gen AI’s ability to generate code can further assist with the transformation.
Legacy modernization is a daunting challenge – it involves a lot of time, financial resources, and effort. Banks should hire trusted financial software development companies that know the ropes to help smoothly transform the existing infrastructure while also providing end-to-end support in building a powerful Gen AI solution.
3. Ethical Challenges
Among the biggest concerns for the banking sector is the potential for AI-based systems to generate outcomes or advice which are biased or unfair.
Key points to consider:
- The resulting outputs can be biased if the data used to train a Gen AI model is incomplete or insufficient. Algorithmic bias may lead to unfair and discriminatory lending decisions for certain population groups.
- Since Gen AI models are complex and sophisticated, bank employees may have a hard time interpreting the output of AI algorithms, or how that output was computed, which would mean that the bank may be unable to explain the reason behind a decision to customers or regulators.
- Gen AI models can confidently produce wrong answers, referred to as “hallucinations”. While looking hyper-realistic, these answers are entirely fictitious, which could lead to catastrophic losses if applied in the context of banking and investments.
To address these issues, it’s critical to integrate human expertise into Gen AI’s decision-making processes every step of the way. Such a human-in-the-loop approach is the only way to reliably detect anomalies before they lead to an actionable decision. Using Gen AI to produce initial responses as a starting point and creating AI-human feedback loops can significantly improve decision making accuracy.
4. Skill Gaps and Talent Acquisition
The talent shortage is another barrier standing in the way of Gen AI adoption in the banking sector. According to John Mileham, CTO at Betterment, “Currently, Gen AI is so new that you can’t really hire a whole lot of experience”.
Integrating Gen AI into banking operations will certainly reshape many roles in the banking workforce in that workers will have to learn new skills or change occupations.
To bridge the skills gap, financial services firms will have to figure out what new skills the workforce will have to acquire and whether they need to reskill and upskill existing employees or hire new ones. This will require extensive investments in retraining and hiring initiatives to meet changing talent needs. Providing internal training programs for employees is key to generating excitement and equipping your existing teams with the resources, skills, and capabilities required for the new roles, such as prompt engineering or model fine-tuning skills.
5. Regulatory Risks
Gen AI is still largely unregulated. This poses a significant barrier to the large-scale adoption of Gen AI in the banking industry.
As the Chief Executive of the UK’s Financial Conduct Authority (FCA) said, “While the FCA does not regulate technology, we do regulate the effect on and use of tech in financial services … With these developments [the growing use of Gen AI], it is critical we do not lose sight of our duty to protect the most vulnerable and to safeguard financial inclusion and access”.
While full regulation of AI by the government is under consideration, the potential value of an extensive application of Gen AI should be balanced against regulatory risks. Fortunately, Gen AI itself provides the finance sector with an efficient means of keeping abreast of changing regulatory environments.
How to Implement AI in the Financial Sector
To successfully implement Gen AI in the banking sector, financial institutions should consider the following five recommendations.
These are key essentials to focus on for a successful Gen AI implementation strategy. However, it is worth noting that they only provide a foundational starting point for building robust Gen AI solutions. Banks will need even more comprehensive implementation roadmaps that detail a wider range of strategic considerations.
1. Define priority areas and set goals
First and foremost, as with any new technology, banks need to have a clear goal that aligns their efforts to business impact. That is, banks need to be clear about why they need Gen AI.
Here are four steps that banks can follow to do this:
- Specify priority areas (functions or units) to experience the biggest impact from Gen AI and plan for specific use cases (e.g. frontline copilot, customer operations, or discovery of regulatory changes)
- Clearly define the objectives
- Examine the interoperability of current data infrastructure with Gen AI tools, assess skills, and evaluate data and technology
- Optimize infrastructure and systems to be capable of supporting Gen AI. A good option would be hybrid infrastructure, which allows banks to work with private models for sensitive data while also leveraging public cloud capabilities.
2. Pilot the technology
Start with a pilot project to evaluate the feasibility of the technology, analyze its potential risks, and measure the adoption.
Train, deploy, and test the Gen AI system on a small scale before expanding it to critical use cases like loan underwriting or generating investment strategies.
Once this is done, a bank should be able to answer the following question: Is the system ready for enterprise-changing initiatives?
3. Collaborate with technology partners
Partnering with technology providers experienced in Gen AI can help banks navigate the complexities of implementation and integration. These partners can offer valuable insights and support throughout the process.
4. Invest in Talent and Training
Building a skilled workforce is essential for successful Gen AI adoption. Invest in training programs for existing employees and attract new talent with the necessary expertise.
5. Establish strong controls
Given that Gen AI brings new risks to the financial services industry, banks and financial institutions need to design new governance frameworks and control systems from the very outset, both for internal use cases and third-party tools, to promote the responsible use of the technology. This refers both to unregulated processes such as customer service and heavily regulated operations such as credit risk scoring.
Closing Thoughts
The integration of Gen AI in banking has the potential to transform the sector, yet it is not without its challenges. As banks navigate data security concerns, legacy system constraints, ethical considerations, skills gaps, and regulatory risks, adopting a cautious and strategic approach is paramount.
By prioritizing clear objectives, piloting technology, collaborating with experienced partners, investing in talent, and establishing robust governance frameworks, financial institutions can harness the benefits of Gen AI while mitigating associated risks.
Ultimately, the journey to full-scale Gen AI adoption will be gradual, requiring thoughtful leadership and continual adaptation to realize its potentially transformative impact on the banking industry.
How exactly banks will leverage Gen AI to gain a competitive edge is still not clear and the future will hold its share of surprises. But one thing is clear as banks navigate this new frontier: Gen AI is shaping the future of banking.
Jason Oh is a Senior Manager at TD Bank’s Enterprise Strategy team. Previously, he was at Strategy&, EY, and Novantas as a strategy consultant, where he led / contributed to a range of high-impact projects to improve both top- and bottom-line across the financial services sector.
Image: DALL-E
🔴 Interested in consulting?