Hi, I'm Tak@. I usually work as a system integrator, involved in various system development projects, and in my free time, I enjoy developing web services using generative AI.

In this column, I'd like to discuss how the trending topic of AI is being used in the financial world and what its future holds, all from a perspective you can easily relate to.

How AI is Changing Finance

The financial industry has experienced many transformations throughout its history.

In the past, methods for assessing financial crime risks largely relied on fixed models and manual verification.

When transaction volumes surged, it became time-consuming and labor-intensive, often leading to a high number of false positives (incorrect alerts).

This is where the power of AI comes into play.

AI, by leveraging technologies like machine learning, natural language processing (NLP), and predictive analytics, can now real-time identify patterns and unusual activities that were previously undetectable using traditional methods.

For instance, one financial institution, by feeding AI customer transaction records and customer information (such as age and occupation), was able to quickly detect unusual cross-border money movements that had gone unnoticed before.

This is incredibly helpful because AI not only "finds" early signs of financial crime but also assists in determining subsequent countermeasures.

AI's impact on the financial industry brings five significant benefits:

• More Accurate Fraud Detection: It can analyze vast amounts of data from various angles, identifying new fraud schemes and unexpected risk patterns much faster.
• Reduced False Positives: By decreasing the numerous false alarms generated by traditional methods, it alleviates "alert fatigue" for audit and compliance officers.
• Continuous Learning: AI keeps learning from new examples, improving its fraud detection capabilities over time.
• Constant Monitoring: It can instantly review daily transactions and customer activities, enabling proactive responses before problems escalate.
• Improved Audit Planning: AI helps streamline audit planning by identifying high-risk areas.

Previous risk assessment methods tended to rely on static models that were used indefinitely.

Many organizations continued to use rules and scoring systems created with data from years ago, without much reconsideration.

However, introducing AI allows for flexible adaptation, breaking free from these "old ways."

The Forefront of AI Utilization in the Financial Industry

AI is truly beginning to be widely adopted in the financial world.

It's playing crucial roles, especially behind the scenes, supporting daily operations, streamlining customer interactions, and safeguarding companies.

Streamlining Internal Operations

Financial institutions often deal with a large volume of documents and operate complex systems.

AI significantly assists with these "white-collar" desk jobs.

For example, AI can help in drafting contracts, meeting minutes, and internal inquiry responses.

It's also active in system development, automatically generating programming code and devising system test items.

Major financial institutions like Mizuho Financial Group, Kyoto Bank, and SMBC Group are all implementing such AI assistant tools.

That reminds me, I once struggled to accurately replicate MOVE statements during a straight conversion of a COBOL system to C#.

It would have been much easier with AI!

OCR (Optical Character Recognition) technology, which converts documents into digital text, and information retrieval, which quickly finds necessary information, are also areas where AI excels.

Given the massive volume of documents handled by financial institutions, improving the efficiency of these tasks drastically reduces manual labor.

Transforming Customer Interactions

AI is also significantly changing how financial institutions interact with their customers.

You might have already experienced chatting with a chatbot on a website or interacting with an automated voice system over the phone.

AI-powered chatbots and voice AIs can now answer customer queries 24/7.

This allows customers to get their questions resolved without waiting, and financial institutions can reduce their human resource burden.

Furthermore, AI is adept at recommending personalized financial products and services tailored to each customer's specific situation.

For example, services like Monex Securities' "AI Hatch," SBI Securities' "SBI Wrap AI Investment Course," and Rakuten Securities' "Investment AI Assistant" use AI to analyze market trends and customer information to offer investment ideas and asset management proposals.

In the insurance industry, Dai-ichi Life is using AI to review and suggest optimal insurance plans to customers, and Tokio Marine & Nichido is offering combined life and non-life insurance proposals.

Smartphone apps for insurance consultations and AI matching services that introduce customers to compatible sales representatives are also emerging.

Enhancing Corporate Risk Management

In the financial world, protecting the company and its customers from money misuse and fraud is paramount.

AI is demonstrating remarkable capabilities in this area of risk management.

AI monitors financial transactions and quickly identifies unusual activities.

This includes, for example, monitoring transactions to prevent money laundering (the act of making illegally obtained money appear legitimate) and cross-referencing against lists of internationally sanctioned individuals.

It also helps detect fraud early by identifying unusual frequencies, times, or amounts from credit card usage history and online banking login patterns.

NEC provides services that use AI for fraud and risk detection, and companies like Daiwa Securities and SBI Securities have adopted these to strengthen their systems for identifying unfair transactions.

Overseas financial institutions also offer many examples of leveraging AI for risk management: DBS Bank uses AI to detect fraudulent transactions, Mastercard detects fraudulent transactions in milliseconds, and Lemonade, an insurance company, pays out bicycle theft claims in just 2 seconds using an AI chatbot.

Domestically, Japanese banks are also making strides: Chiba Bank has implemented an AI system to detect fraudulent transfer scams and unauthorized account usage, Sumishin SBI Net Bank has developed its own AI monitoring system for unauthorized transfers, and Seven Bank uses AI to predict discrepancies in ATM cash deposits and withdrawals.

These are examples of how AI acts as the "eyes" of financial institutions, identifying hidden fraud and risks.

New Horizons AI Brings to Finance

The impact of AI on the financial industry isn't limited to what we currently see.

It holds the potential to significantly transform the future of financial services and the value we can receive.

Increased Efficiency and Quality

The adoption of AI primarily brings about increased efficiency and quality.

Tasks that previously required manual review of massive amounts of documents and transaction data were inherently time-consuming, labor-intensive, and prone to oversight.

However, with AI filtering data and proactively identifying unusual values or risk patterns, people can now focus on deeper investigations and customer interactions.

AI is significantly accelerating operations by improving fraud detection accuracy, reducing false positives, and providing real-time monitoring.

This, in turn, helps financial institutions cut operational costs and cybersecurity expenses.

As Japan faces a declining population and a decrease in per capita productivity, the increase in productivity driven by AI holds immense importance.

Creating New Value

AI not only streamlines existing operations but also has the power to create new services and value for customers that were previously unimaginable.

By analyzing vast amounts of financial and customer data, AI can suggest truly personalized financial products and services for each individual.

Furthermore, AI's ability to analyze market trends in detail and predict the future will enable smarter decisions in asset management and investment.

For example, the French insurance company AXA offers "telematics insurance" using AI.

This involves AI analyzing car driving data (speed, sudden braking, etc.) and offering reduced premiums to safe drivers.

This provides customers with fair and transparent premiums and encourages safer driving habits.

Overseas brokerage firms are finding "fully automated trading" powered by AI to be a powerful tool.

AI monitors the market 24/7 and automatically executes trades at optimal times.

Since it's not influenced by human emotions and doesn't miss opportunities, investment efficiency increases, aiming for more stable results.

Another key feature is "advanced risk management," where AI automatically adjusts risk levels according to market conditions.

Important Considerations When Working with AI

While AI offers significant benefits, there are many crucial points to consider when using it.

The "Why?" Barrier

AI, especially what's known as "generative AI," often acts like a "black box" where it's difficult for humans to understand how it arrived at its conclusions.

Financial institutions have a responsibility to explain their judgments and actions to customers and regulatory bodies.

If an AI-generated outcome leads to a disadvantageous decision for a customer, it's crucial to explain "why that decision was made." This can be challenging due to AI's complexity.

Furthermore, AI carries the risk of "bias," where it may produce discriminatory results if the data it learns from is skewed.

For instance, if AI is used for loan applications, it might make decisions that are unfavorable to people of a specific gender or region.

To address this, it's necessary to implement processes where humans always review AI's decisions and to proactively inform customers about the possibility of AI generating inaccurate information.

The Financial Services Agency (FSA) and the Bank of Japan (BOJ) also emphasize the importance of "explainability" and "fairness" and continue to engage in dialogue on these issues.

Data Quality and Information Management

AI is highly dependent on the quality of the data it's trained on.

If it learns from outdated or biased data, AI will also make biased judgments.

Therefore, to use AI effectively, it's crucial to prepare a large amount of high-quality data and manage it appropriately.

Moreover, customer personal information and company confidential information handled by financial institutions are highly sensitive.

When inputting this information into AI, there's a risk of unintended data leakage.

When using cloud-based AI services, it's essential to ensure that data isn't used for retraining and that mechanisms are in place to prevent external leaks.

The potential for employees to use external AI services without company permission, leading to "shadow IT," also increases the risk of information leakage.

Security and Misuse

The evolution of AI also raises concerns about increasingly sophisticated cyberattack methods.

For example, there might be a rise in phishing scams using highly realistic emails or fake websites created by AI to deceive us.

"Deepfake" technology, which allows for the easy creation of fake videos and audio of specific individuals, also poses a risk of increasing fraudulent activities.

In the past, a fake image of an explosion near the Pentagon circulated on social media, temporarily causing stock prices to drop.

There's also a risk that financial institutions' identity verification systems could be breached using such fake images or videos.

To counter these risks, it's crucial for companies to strengthen overall security measures, provide thorough employee training, and establish a system for responding to risks by collaborating not only with AI specialized departments but also with legal and compliance departments.

Impact on the Financial System

The widespread adoption of AI could lead to a "herd effect," where investors and financial institutions act similarly based on AI's judgments, causing the market to move significantly in one direction.

If the provision of AI services becomes concentrated in a few companies, this impact could be even greater.

Bank of Japan Governor Ueda stated that while "equations can be decided by AI," he emphasized the importance of "making the best judgments as humans," suggesting that there are aspects that cannot be decided solely by AI and where human judgment remains crucial.

It's vital to view AI merely as an aid, and the ultimate responsibility should rest with humans.

The Future Path of Finance and AI

AI is still in its early stages of evolution, and its full-scale utilization in the financial industry is yet to come.

Preparing Governance and Fostering Talent

To safely use AI, it's indispensable to establish robust rules and frameworks within the company (known as "AI governance").

Given the rapid pace of AI evolution, simply creating rules once isn't enough; they need to be flexibly reviewed and revised in response to changing circumstances.

Furthermore, there's still a shortage of personnel with the specialized knowledge to effectively utilize AI.

It's necessary to cultivate AI experts within the company or collaborate with external experts.

The Bank of Japan and the Financial Services Agency also emphasize the importance of human resource development for AI utilization and public-private, international cooperation.

Steady Steps Towards Adoption

It's advisable to introduce AI incrementally.

First, clearly define the purpose for using AI (e.g., anti-money laundering or fraud detection).

Next, establish mechanisms for preparing and managing the data that AI will operate on.

Then, conduct "pilot projects" where AI is used within a limited scope, and gradually expand its application based on the results.

When developing an AI system, there's a choice between developing everything in-house or using external services.

The best option depends on the company's objectives, desired functionalities, and security stringency.

When introducing new AI services, it's crucial to thoroughly discuss their usability (UI/UX) and how they will be integrated into existing workflows (operational processes) beforehand.

Moreover, once AI-powered services are launched, collaboration across various departments is necessary for aspects like personal information handling and system infrastructure. Therefore, establishing a system for smooth inter-departmental communication is also important.

The Bank of Japan and the Financial Services Agency are also considering leveraging AI for tasks like analyzing prices and interest rates, and supervising financial institutions.

They have indicated a flexible approach to reviewing existing rules and guidelines, acknowledging the rapid advancements in AI technology.

A Collaborative Future Between Humans and AI

AI should not be seen as something that will take away our jobs, but rather as an "excellent auxiliary auditor" that expands our human capabilities.

AI, with its ability to store vast amounts of information and retrieve necessary data instantly, is like a highly capable junior auditor.

However, instead of simply accepting AI's output, it remains crucial for humans to make the final judgments.

By fostering cooperation between AI and humans, leveraging each other's strengths, we can achieve better outcomes and further advance the financial industry.

The future of finance and AI hinges not just on technological progress, but on how we approach AI and how we collaborate with it.

We look forward to this new collaboration enriching our lives in the future.

Follow me!