More Stories

Informist, Monday, Apr. 13, 2026

 

Please click here to read all liners published on this story

--RBI Swaminathan: AI-enabled systems can simplify customer interactions
--RBI Swaminathan:AI-enabled systems more intuitive, responsive for customer 
--CONTEXT:RBI Deputy Governor Swaminathan's remarks at event in Thanjavur Sat
--RBI Swaminathan: AI can help improve credit delivery 
--RBI Swaminathan:AI can identify borrowers who may otherwise remain excluded 
--RBI Swaminathan: AI can contribute to fraud detection, risk management 
--RBI Swaminathan: AI risk mgmt qualities especially important in payments
--RBI Swaminathan: There is role for AI in compliance, supervision 
--RBI Swaminathan: Without safeguards, AI can amplify existing weaknesses
--RBI Swaminathan: Without safeguards, AI can create new forms of harm
--RBI Swaminathan:AI use may lead to bias,unfair outcomes from training data
--RBI Swaminathan: Advanced AI systems opaque, work like black boxes
--RBI Swaminathan: AI use raises concerns on data privacy, misuse
--RBI Swaminathan: Data governance in AI cannot be treated as a side issue
--RBI Swaminathan: AI use poses model risk, concentration risk, cyber risk
--RBI Swaminathan:AI use may turn individual into correlated vulnerabilities
--RBI Swaminathan: Humans, institutions must be accountable for their AI use
--RBI Swaminathan: Institutional capacity must be stronger if AI being used
--RBI Swaminathan: Financial inclusion must be design objective of AI use

 

MUMBAI – Reserve Bank of India Deputy Governor Swaminathan J. laid down guiding principles on the use of artificial intelligence in the financial sector in a speech on Saturday. He laid out opportunities and risks of the use of the emerging technology, but underlined that humans must be responsible for its use.

 

Swaminathan, who heads the RBI departments for supervision and financial inclusion, said AI could support decision-making at banks and non-banking financial companies, but the institutions cannot outsource responsibility to an algorithm, vendor or platform. The system must also be fair and explainable to customers affected by the decisions taken while using AI-assisted models.

 

The senior RBI official flagged concerns about the opacity of advanced systems that operate like black boxes for customers, managers at financial institutions, and even regulators. While expanding credit to borrowers that might otherwise remain excluded, the credit assessment of AI systems could create outcomes that are difficult to justify and harder to detect, he said. 

 

"The enduring task, therefore, is to make finance more intelligent, without making it less human; to make it more digital, without making it less accountable; and to make it more inclusive, without making it less prudent," Swaminathan said while delivering the V. Narayanan Memorial Lecture at SASTRA University in Thanjavur, Tamil Nadu, on Saturday.

 

As the technology component becomes more complex, supervisors need to be confident that systems are robust, auditable and well-governed, Swaminathan said. Managements should also understand the limitations and assumptions of the model to be able to explain the basis of important decisions and convey it to the customer. AI-enabled systems can make customer interaction simpler, more intuitive, and responsive. However, a decision that materially impacts a citizen's economic life cannot be defended by saying "the machine decided", the deputy governor said. 

 

Swaminathan also expressed his concerns about data privacy and its misuse. He said data-governance couldn't be treated as a side issue. Financial institutions must think seriously about consent, storage, sharing access controls and purpose limitations, and not about privacy and innovation as mutually opposed.

 

AI systems can also meaningfully add to fraud detection and risk management, and flag suspicious activity that can be acted upon quicker by regulated entities. These are especially important in the case of payments infrastructure, for which convenience and safety are both primary goals.

 

AI systems also pose a concentration risk and have risks associated with the models used, Swaminathan said. The models may be trained on data with biases as the data carries the imprint of past behaviour, which may then get embedded and reproduced in greater scale. A flawed AI model could now affect decisions across millions of customers, unlike earlier times when the risk was concentrated to limited accounts by decisions made by humans. 

 

"Further, if multiple institutions rely on similar models, common datasets, a small set of vendors or shared infrastructure, individual vulnerabilities can become correlated vulnerabilities," Swaminathan said. "This is where even a local weakness can acquire broader systemic significance."

 

Use of the technology also exposes banks to newer forms of cyber risk, since AI can equip attackers, the RBI official said. Earlier this month, AI company Anthropic launched a model that it deemed too powerful to release publicly, as it was able to identify security vulnerabilities across most webpages on the internet. Swaminathan said fraudsters could use AI to craft more convincing phishing attempts and automate malicious activity.

 

The AI-enabled system plays a role in compliance and supervision and cannot rely on periodic reporting and backward-looking assessments, he said. The use of AI in the financial sector poses governance, capability, and cultural challenges along with technological challenges and in order to tackle these, board members and senior managements must be equipped to ask the right questions, he said, adding that institutions could manage risks more effectively with the help of intelligent tools.

 

The deputy governor said that the best innovation was the one that made formal finance more inclusive, safer and simpler, not the one "that dazzles those already well served", and that "inclusion should be innovation's highest purpose". Swaminathan, in his closing remarks, said AI-led systems could serve a noble cause if they help widen opportunity, improve trust, and protect customers.  End

 

IST, or Indian Standard Time, is five-and-a-half hours ahead of GMT

 

Reported by Janwee Prajapti

Edited by Avishek Dutta

 

For users of real-time market data terminals, Informist news is available exclusively on the NSE Cogencis WorkStation.

 

Cogencis news is now Informist news. This follows the acquisition of Cogencis Information Services Ltd. by NSE Data & Analytics Ltd., a 100% subsidiary of the National Stock Exchange of India Ltd. As a part of the transaction, the news department of Cogencis has been sold to Informist Media Pvt. Ltd.

 

Informist Media Tel +91 (22) 6985-4000

Send comments to feedback@informistmedia.com

 

© Informist Media Pvt. Ltd. 2026. All rights reserved.