Print

Tete-a-tete around AI in banking sector often begins with algorithms

Published: Aug 03, 2026

By CA Rishabh Swansukha

IT should begin with trust.

For decades, banks competed on branch networks, interest rates and digital convenience. Today they compete on intelligence. AI is rapidly moving from customer support to credit underwriting, fraud detection, treasury management, compliance automation, wealth advisory and autonomous financial operations.

Yet beneath every AI model lies something far more valuable than compute.

Data.

And beneath every dataset lies something even more valuable.

Human trust.

This distinction matters because data is not merely an economic resource. It represents fragments of human lives, our financial aspirations, medical histories, business decisions, spending habits, family responsibilities and increasingly, our behavioural patterns. AI has given institutions an unprecedented ability to derive meaning from these fragments. The challenge is ensuring that this capability remains aligned with public interest.

The next era of banking will not be defined by artificial intelligence alone.

It will be defined by responsible intelligence.

The Ramayana's Lesson on Information

Indian civilization has always understood that information is power.

One of the most fascinating parallels comes from the Ramayana.

Ravana possessed extraordinary strength, unmatched scholarship and technological superiority within Lanka. Yet his defeat became possible when Vibhishana crossed over and revealed strategic knowledge regarding Lanka's military capabilities and Ravana's vulnerabilities.

Viewed through a contemporary lens, one might describe this as history's most consequential strategic information disclosure.

Was it a data leak?

In one sense, yes.

Did it violate loyalty?

Certainly.

Was it unethical?

The answer depends on the moral framework one adopts.

Vibhishana did not disclose information for personal enrichment or commercial gain. He acted because he believed that dharma demanded intervention against injustice.

The episode offers a profound distinction that remains relevant in the AI era.

Not every disclosure of information is harmful.

The intent, proportionality and societal outcome matter.

Modern governance recognizes similar principles. Whistle blowers exposing corruption, financial crimes or systemic fraud often reveal confidential information to protect the public interest. Their actions may technically involve disclosure, yet society frequently regards them as essential to accountability.

The ethical question, therefore, is not whether information moved.

It is why it moved, how it moved and whom it ultimately served.

Technology changes.

The philosophy of trust does not.

The Two Faces of Data Sharing

Public discourse often treats every data leak as equally dangerous.

Reality is more nuanced.

There are fundamentally different categories of data movement.

The first destroys trust.

This includes identity theft, ransomware, insider trading, unauthorized surveillance, financial fraud, manipulation of vulnerable consumers and commercial exploitation without informed consent.

These are violations of both law and ethics.

The second protects society.

Financial intelligence shared to prevent money laundering, fraud detection across institutions, cyber security threat intelligence, epidemiological data used during public health emergencies, or disclosures exposing corruption all represent instances where limited, accountable sharing serves a larger public good.

The difference lies not in the movement of information but in governance.

Responsible AI requires distinguishing legitimate public-interest sharing from exploitative extraction.

The future will demand precisely this sophistication.

Banking is Becoming an Intelligence Industry

Historically, banks managed money.

Tomorrow they will increasingly manage decisions.

AI systems already assess creditworthiness, detect anomalies, personalize investment strategies and monitor compliance obligations at scales impossible for human teams alone.

Agentic AI promises an even more autonomous future, where intelligent systems negotiate loans, optimize treasury positions, monitor regulatory obligations and interact directly with customers.

These capabilities create extraordinary productivity.

They also multiply systemic risks.

A compromised AI system no longer leaks static customer records.

It can infer future intentions.

It can predict vulnerabilities.

It can manipulate behaviour.

It can automate financial deception.

The nature of cyber risk therefore evolves from theft to influence.

That distinction should concern policymakers as much as technologists.

Data is No Longer the Asset

For years, executives repeated a familiar phrase:

"Data is the new oil."

That analogy has outlived its usefulness.

Oil creates value through extraction.

Data creates value through trust.

Unlike oil, data belongs to people before it belongs to institutions. It carries identity, dignity and agency. Its misuse harms individuals long after the immediate breach.

A bank suffering a data breach loses more than records.

It loses confidence.

And confidence remains the fundamental currency of finance.

Technology scales capability.

Trust scales adoption.

The Emerging Threat of Inference

Future privacy risks may not arise from stolen databases.

They may arise from conclusions.

AI increasingly infers characteristics that individuals never explicitly revealed-health conditions, political preferences, financial stress, purchasing intentions, or emotional states. These inferred attributes can shape lending, insurance, pricing and even employment opportunities.

This shifts the privacy debate from What data do organizations hold? to What can they infer?

That distinction is critical for regulators. Protecting explicit data alone may not be sufficient if AI systems can reconstruct highly sensitive profiles from seemingly harmless information.

The Responsibility of Financial AI

Responsible AI in finance extends well beyond model accuracy.

It requires institutional discipline.

Banks must ensure that AI systems remain explainable, auditable and accountable. Human oversight should accompany high-impact decisions, particularly those affecting credit access, insurance eligibility or financial inclusion. Privacy-enhancing technologies, data minimization, purpose limitation and robust cyber security should become foundational design principles rather than compliance afterthoughts.

Governance must evolve from annual policy reviews to continuous oversight, where AI systems are monitored throughout their lifecycle. Equally important is educating customers about how AI is used, what data is processed, and what rights they possess.

Regulating Abuse Without Hindering Innovation

The challenge before regulators is delicate.

Excessive restriction may discourage innovation.

Insufficient regulation may encourage exploitation.

The objective is not to regulate AI itself.

It is to regulate harmful outcomes.

This distinction deserves far greater attention.

An AI model predicting crop yields is fundamentally different from an AI system manipulating vulnerable borrowers into unsuitable financial products.

Similarly, personalized education differs profoundly from algorithmic discrimination.

Regulation should therefore become risk-based rather than technology-based.

The higher the societal impact, the higher the governance obligations.

India's DPDP Act is an important milestone, but the AI era will demand complementary frameworks addressing algorithmic accountability, automated decision-making, explainability, auditability and liability.

The future belongs not merely to digital governance but to intelligent governance.

The Orange Economy Needs Protection Too

Data governance discussions often focus on consumers.

Creators deserve equal attention.

The Orange Economy-comprising artists, writers, designers, filmmakers, musicians, educators, software developers and other creators-produces the intellectual and cultural assets that increasingly train AI systems. Yet many creators face an uneven distribution of value. Their work fuels innovation, while compensation, attribution and bargaining power remain fragmented.

Rather than an "unfair distribution mechanism," what is needed is a fair value-distribution mechanism -one that rewards creative labour when it contributes to AI-enabled products and services. Such a framework could combine transparent licensing, provenance tracking, collective bargaining where appropriate, and technological tools that identify when protected works are used. It should also preserve space for lawful exceptions, public-interest research and education, ensuring that innovation is not frozen while creators retain meaningful rights.

Creativity is not merely content.

It is national intellectual capital.

If AI becomes one of the largest consumers of human creativity, economic policy must ensure that creators participate in the value they help generate.

India's cultural diversity could become one of its greatest AI advantages-provided the incentives remain fair.

The AI Bharat Framework: From Data Economy to Trust Economy

India's next leap should not simply be becoming the world's largest AI market.

It should aspire to become the world's most trusted AI economy.

This requires what I describe as the AI Bharat Framework, built upon five interconnected pillars:

Together, these pillars move the conversation from compliance to confidence.

From Financial Capital to Trust Capital

For centuries, economists measured wealth through land, labour and capital.

The digital age added data.

The AI age introduces a new asset.

Trust Capital.

Institutions that accumulate Trust Capital will attract customers, investment and long-term resilience. Those that erode it may find that no amount of computational sophistication can compensate for the loss.

The lesson from the Ramayana is not merely that information can defeat power.

It is that information acquires moral meaning through purpose.

Vibhishana's disclosure served dharma rather than domination. By contrast, data theft for fraud, manipulation or exploitation serves only private gain at public cost. Modern governance must preserve this distinction. It should protect citizens from misuse while enabling accountable disclosures that advance justice, financial integrity and societal well-being.

Artificial Intelligence is transforming banking.

But banking has always been built on something that no algorithm can manufacture.

Confidence.

And confidence begins where responsible stewardship of data meets ethical judgment.

CA Rishabh Kumar S is a Chartered Accountant, author of The Networker, Founder of AI Champion Hub, and an AICA Level 2 certified AI professional. He writes about artificial intelligence, enterprise transformation, taxation, digital public infrastructure, and the future of India's economy. His work focuses on bridging technology, policy, and business to create inclusive and scalable innovation. Views expressed are personal. Reach out to him rishabh@bizstreet.biz

(Disclaimer: The views expressed are strictly personal. The author confirms that no AI has been used to generate this content. )