So far, financial institutions are mostly using AI to improve the customer experience. But as Alex Roberto of Phaxis argues, there may be some good reasons why banks and other institutions aren’t yet embracing AI — their systems and technology won’t let them.
The potential for artificial intelligence to revolutionize our lives and enhance experiences is well-documented. Many of us can attest to this from our encounters with smart-home devices, virtual assistants and chatbots or personalized recommendations from companies like Amazon and Netflix. Similarly, consumers of financial services have enjoyed the benefits of AI as institutions have rolled it out to optimize service operations, detect fraudulent activity and provide tailored solutions to their customers.
Despite the significant impact of these technological advancements on our daily lives, the use of AI in the global fight against money laundering remains woefully inadequate. While many vendors offer solutions to reduce false-positive alerts, improve efficiency and enhance visibility into the effectiveness of operational processes and risk mitigation strategies, poor data quality within many banks hinders the widespread adoption of AI to combat money laundering. (Even global behemoths like Google are getting in on the action.)
Still, some underlying issues stand in the way of progress:
- Aging technology: Many financial institutions still rely on legacy systems that are not equipped to handle the demands of the digital era. These outdated systems were developed using programming languages and hardware platforms that are now obsolete, making them challenging to maintain and update. Their design also often lacks modularity, which makes the addition of new functionality or services costly and difficult. Given that AI relies on high-quality data to train and improve its algorithms, outdated systems serve as a significant barrier to the widespread adoption and use of AI to fight financial crime.
- Disparate systems: The above-referenced legacy systems were often created in silos for specific departments or activities within a given institution, making it difficult to integrate them. This lack of integration leads to inefficiencies and inconsistencies that negatively impact data quality and customer experience. The issue is further compounded for banks that have grown through mergers and acquisitions or by adding new lines of business, often resulting in a new set of systems that cannot be easily integrated into the bank’s core systems. These challenges pose obstacles to effective AML efforts since they hamper an institution’s ability to optimize its operations and ultimately provide high-quality data for AI-powered AML solutions.
- Ineffective controls: Poor data governance and controls can compromise the quality of data used by AML systems. Incomplete or inaccurate data inputs can arise from issues in managing and controlling data collection, storage, use, sharing and protection at financial institutions. Disparate systems containing financial transaction data without standardization will negatively impact pattern recognition accuracy, resulting in false positive alerts. Furthermore, inadequate data governance may lead to privacy violations and non-compliance with AML regulations that mandate strict confidentiality around financial information.
Businesses Continue to Embrace AI for Hiring, Customer Service & Cybersecurity
About six months from the splashy launch of Open AI’s ChatGPT and despite continued global regulatory concerns about AI’s proliferation, new research indicates that corporations are using the technology across their operations and are likely to keep expanding deployment of AI.
Read moreHow can banks make the most of AI?
With these significant challenges and lots of hard work ahead, let’s discuss the steps that financial institutions need to take to address these issues and ultimately modernize their AML efforts:
Gain detailed insight into IT infrastructure
To address aging technology, financial institutions need to conduct a detailed analysis of their IT infrastructure. Defining organizational goals — such as enhancing system performance, improving security or reducing cost — is the first step in this process.
This allows the institution to conduct thorough assessments around system vulnerabilities, application architecture, data flows and access controls. Once the gaps are understood and prioritized, actionable steps can be identified to optimize system performance, mitigate risks and ensure compliance with regulatory requirements.
As part of this process, financial institutions should consider migrating to cloud-based services. This allows for the pivot from physical data centers and hardware infrastructure toward a more flexible and scalable model. It also allows the institution to take advantage of templates for network architecture, security policies, compliance certifications and more.
For those who are already in the process of migrating into the cloud, it’s important to prioritize changes that will improve the client experience. These changes can include enhancements to online banking and mobile apps, improving security/client authentication and automating due diligence and data verification methods to mitigate risk and meet compliance requirements.
Integrate systems
Once an institution has migrated to the cloud, banks grappling with disconnected legacy systems will be in a strong position to implement an extensive system integration program. This program would involve the consolidation of disparate systems into a unified, seamless framework via the development and deployment of robust application programming interface (API) solutions.
The resulting integrated system would standardize and streamline important functions like transaction processing, customer relationship management and risk analysis. Further, the retirement of obsolete legacy systems would lead to cost savings and reduce maintenance overhead. This will bring the institution one step closer to streamlining the delivery of standardized data which will allow AML-focused AI tools to be used to their maximum potential.
Get serious about data governance
Effective data governance practices are a must to mitigate poor data management and controls within financial institutions. Establishing a framework for data governance enables organizations to set guidelines and policies around data collection, storage, use, sharing and protection.
This includes regular audits of systems and process workflows to verify the accuracy of data inputs. In addition, adopting standardized processes for handling confidential information will also prevent privacy breaches and ensure compliance with AML regulations.
Implementing these measures will improve internal efficiency while creating a more robust regulatory environment for financial institutions. Most importantly, effective data governance means consistent, high-quality data that allows for the effective deployment of AI to analyze red flag alerts in real-time while automating due diligence and verification methods to the greatest extent possible.
It’s not up to banks alone
Governments and regulators must also play a key role by aligning their requirements with industry-wide standards for data management and governance and by facilitating secure information sharing between financial institutions.
Specifically, regulatory requirements should be updated to meet the Wolfsberg Group principles for the use of AI in AML, which emphasize legitimate purpose, proportionate usage, technical expertise, accountability and oversight, and openness and transparency. This leadership and guidance is monumentally important to ensure that progress is made collectively, as many institutions will otherwise prioritize projects and issues according to their own unique circumstances. To further support the implementation of these action plans and ensure they are performed confidently, governments and regulators should consider taking the following steps:
- Providing tax benefits and other incentives to institutions that prioritize these plans.
- Allocating funds towards research and development of new technologies and methods to improve systems and data security.
- Use regulatory authority to drive secure, efficient data management leading to effective use of AI in AML efforts.
- Ensure government contracts are assigned to firms demonstrating dedication to improvements along these lines.
A new era has arrived, and it’s time for institutions and governments to unite in a more powerful and collaborative effort to combat money laundering. As we hear impassioned calls for a deeper conversation about AI regulation, we have a unique opportunity to implement measures that protect our citizens while safeguarding the integrity of our financial systems worldwide. Together, we can create a more secure, stable and prosperous world for generations to come.
Alex Roberto serves as managing director and solutions lead for the professional services division of Phaxis. In this role, he advises the firm’s financial institution clients on how to tackle their most pressing challenges in the battle against financial crime. As a 20-year veteran of the banking industry, Alex has experience and expertise that spans across AML/KYC compliance, strategy and operations, risk management, process improvement, business development and leading large and complex business units within top 10 U.S. banks. Through these experiences, he has led multiple efforts to scale operations by streamlining processes, implementing automation, redesigning operating models and delivering on C-level strategic initiatives.
