Smart Banks: AI, Pros and Cons

AI is revolutionizing smart banking. Discover banking industry pros and cons, including efficiency, personalization, and ethical/security challenges.

The Silent Revolution Behind the ATM: How Artificial Intelligence is Transforming the Face of Banks, Between Promises of Efficiency and Risks of Discrimination

Imagine walking into a bank and being greeted by an advisor who knows you better than your family: they know exactly when you get your paycheck, predict your future expenses, and suggest personalized investments before you even ask. This advisor never gets tired, never takes a lunch break, and never has a bad day. It's called artificial intelligence, and it's already changing the way banks operate.

Banking AI is not science fiction: banks like JPMorgan Chase and Bank of America have developed advanced virtual assistants, like Erica, a chatbot that provides personalized financial advice. But behind these innovations lie profound questions: what happens when algorithms decide who deserves a loan? What if artificial intelligence inherits our deepest biases?

What is Banking AI and Why It's Changing Everything

Artificial intelligence in the banking sector is a set of technologies that enables computers to perform a variety of advanced functions, including the ability to analyze data, formulate recommendations, and make decisions that directly impact the financial lives of millions of people.

According to a survey by ABI Lab, 51% of Italian banks use Artificial Intelligence, with applications ranging from security management (48%) to customer service via chatbots (43%), to optimizing internal operations (29%). It is no longer an experiment, but an established reality that is redefining the entire sector.

Banking AI manifests in three main areas:

Intelligent Automation: RPA (Robotic Process Automation) technologies allow for the automation of administrative operations like document processing and customer request management, reducing wait times and improving productivity.

Predictive Analysis: Algorithms can predict future behaviors by analyzing huge amounts of historical data, from spending habits to default risk.

Intelligent Interface: Chatbots and virtual assistants that offer 24/7 support, like Bank of America's Erica, a virtual assistant that, in addition to answering customer questions, is also capable of learning from interactions thanks to deep learning and sends notifications to customers, suggesting areas where they can save money.

The Advantages Revolutionizing Banking

Enhanced Security Against Fraud

Using machine learning algorithms, banks are able to analyze millions of transactions in real-time and identify suspicious activity. A concrete example? Deutsche Bank's "Black Forest" artificial intelligence system analyzes transactions and flags suspicious cases. For each capital movement, several criteria are examined: amount, currency, destination country, and transaction type.

The results are tangible: thanks to this AI application, several cases of fraud and tax evasion have already been uncovered, including one related to organized crime and money laundering.

More Accurate and Inclusive Credit Scoring

AI is revolutionizing credit assessment. Banks use AI to evaluate credit risk more accurately and quickly than traditional scoring models. By analyzing customers' historical and behavioral data, machine learning models can predict the likelihood of a customer defaulting on their debt.

According to a recent study by Banca d'Italia, the use of artificial intelligence by banks allows for more effective processing of quantitative information and can reduce the information asymmetry between intermediaries and their clients.

Service Personalization

AI enables banks to offer a personalized experience to customers by analyzing their transaction data, preferences, and behaviors. A bank can suggest a specific financial product to a customer based on their spending habits or propose a targeted investment plan.

Banks use Artificial Intelligence systems to analyze customers' spending habits and offer personalized advice on investments, savings, and lines of credit.

Unprecedented Operational Efficiency

The automation of routine and time-consuming activities, such as document processing, data entry, and compliance checks, reduces manual workload, minimizes errors, and lowers operational costs.

A practical example: some banks have implemented AI-based systems to automatically analyze financial documents, speeding up loan processing and increasing responsiveness to new requests.

Hidden Risks: When the Algorithm Discriminates

The Problem of Algorithmic Bias

Algorithmic bias occurs when a model systematically produces skewed results to the detriment of specific groups. In the credit sector, this can mean the unjustified rejection of applications from individuals belonging to minorities, or the assignment of worse terms and conditions.

A prime example is Amazon, which had to scrap its AI-based recruitment system because it systematically discriminated against female candidates. The algorithm had "learned" from historical data that men were more frequently hired for technical roles.

In the banking sector, this translates into invisible but real discrimination. All major financial institutions possess vast datasets on customers' financial profiles which, when integrated with machine learning operations, can produce bias.

Privacy and Data Control

AI is based on the massive collection of personal information, and banks must guarantee the protection of customer data, complying with regulations like the GDPR and preventing potential breaches and inappropriate uses.

An advanced algorithm or an AI virtual assistant would be perfectly capable of deciding whether a customer is eligible for a loan or to receive a specific commercial offer. However, this type of automation conflicts with Article 22 of the GDPR, which explicitly prohibits companies from making automated decisions that produce significant legal effects.

Transparency and "Black Boxes"

One of the most pressing problems is the lack of transparency. Transparency is particularly important in sectors where AI makes decisions that have a significant impact on people's lives, such as finance. If an AI system rejects a bank loan, it is crucial that the affected individuals have the right to understand the decision-making process.

The need for XAI (Explainable Artificial Intelligence) stems from the fact that many advanced AI models are often considered "black boxes." This means that, although they are capable of producing highly accurate results, the process through which they arrive at these conclusions remains opaque.

The Regulatory Framework: The European AI Act

To address these risks, the European Union has approved the AI Act, the world's first comprehensive legal framework on AI. As of February 2, 2025, the provisions of the European Regulation on Artificial Intelligence concerning systems that pose unacceptable risks will be in force.

The growing spread of AI represents both an opportunity and a challenge. On one hand, the AI Act introduces a clear regulatory framework aimed at balancing technological innovation with consumer protection.

For banks, this means:

High-Risk Systems: High-risk AI systems must have a risk management system that iteratively identifies, assesses, and manages potential risks.
Mandatory Transparency: Companies must be able to explain how and why certain decisions are made.
Continuous Monitoring: It is necessary to ensure that AI systems are transparent, with the ability to trace algorithmic decisions.

The Impact on Banking Work

According to a report examining banking giants like Citigroup, JPMorgan Chase, and Goldman Sachs Group, the interviewed Chief Information and Technology Officers stated that on average they anticipate a net reduction of 3% of their workforce.

However, the reality is more complex. Teresa Heitsenrether, who oversees JPMorgan's AI efforts, stated that the bank's adoption of generative AI has so far increased jobs.

The transformation seems to be moving towards a hybrid model of human analysts and AI rather than a simple replacement.

Concrete Examples: Smart Banks in Action

JPMorgan Chase and Machine Learning

COIN, the chatbot launched in 2016 by JPMorgan Chase, demonstrates that even back-office operations can undergo a revolution. The system analyzes legal contracts in seconds, a task that would require 360,000 hours of human work.

DBS Singapore and Total Assistance

DBS Singapore has created a chatbot considered a virtual banking assistant fully available to the customer. Interactions occur via voice or text, and the assistant can anticipate and respond to over 10,000 common questions.

Santander and Voice Recognition

Santander UK has launched a voice recognition technology within their SmartBank app, allowing their customers to manage their savings using only their voice.

Towards a Balanced Future: Human and Intelligent Banks

The challenge is not deciding whether to embrace or reject AI, but how to implement it responsibly. Most models using artificial intelligence react similarly to the incentives provided by users: they are more likely to behave unethically when told that unethical actions will lead to substantial monetary gains, according to a study by Banca d'Italia on AI ethics in the financial sector.

Best practices for responsible banking AI

Diversity in development teams: It is more crucial than ever to involve and collaborate with various stakeholders and engage them in the design, development, implementation, and evaluation of the AI system.
Continuous audits: It is necessary to identify and understand the sources of bias that can influence an AI project and assess the impact and risk of partiality.
Algorithmic transparency: The principle of transparency, a cornerstone of the GDPR and the new AI Act, ensures trustworthy and ethical AI. Essential for the financial sector, it allows for processing data with precision, improving services and security.
Human oversight: Human intervention is crucial to verify that AI systems operate in compliance with principles of fairness, transparency, and accountability.

New opportunities: robo-advisors and the democratization of investing

Robo-advisors offer personalized financial advice at reduced costs. These AI-based financial advisory tools analyze the user's risk profile and financial goals to suggest tailored investment strategies. According to a study by Business Insider Intelligence, by 2025, robo-advisors will manage over $4.6 trillion in assets globally.

Conclusion: A new contract of trust

Artificial intelligence has the potential to make the banking system more efficient, inclusive, and secure. With the adoption of ethical, transparent, and secure practices, AI can become a powerful ally in digital transformation, improving the services offered and strengthening customer trust.

But this transformation requires vigilance. If properly governed, digital lending represents a great opportunity for the sustainability of the financial system. But, above all, it can help bridge the credit gap that still plagues millions of people excluded from the formal banking circuit.

The future of smart banks will not be determined by technology itself, but by our ability to build a new trust contract between humans and machines. A contract where algorithmic efficiency combines with human empathy, where innovation serves inclusion, and where transparency becomes the foundation of every automated decision.

The banking revolution has begun. It is up to us to ensure it is a revolution for everyone.