Banking Central | RBI's AI ambition: Innovation yes, but not at the cost of chaos

AI can be a double edged sword in banking

The Reserve Bank of India's (RBI) latest report into artificial intelligence with its Framework for Responsible and Ethical Enablement of AI (FREE-AI) is a bold step, no doubt. In a country where financial inclusion is still a work in progress, artificial intelligence promises to bridge gaps —think multi-lingual chatbots dispensing loans to rural farmers to algorithms sniffing out fraud in real-time.

The report, with its seven 'Sutras' such as Trust as Foundation and People First, paints a rosy picture of innovation harmonised with safeguards but implementation calls for caution.

Bank employee unions such as the All India Bank Officers' Confederation (AIBOC) are concerned that this framework could unravel into a regulatory nightmare, exacerbating biases, job losses and cyber vulnerabilities in an already fragile banking sector.

AI can be a double-edged sword that can slice through inefficiencies or inflict deep wounds if mishandled. Several RBI surveys and consultations have acknowledged risks — bias in credit scoring that discriminates against low-income groups, opaque “black-box” models defying explainability and amplified cyber threats from adversarial actors.

That said, the framework's recommendations, while advocating shared infrastructure and sandboxes for innovation, seem to treat risk mitigation as an afterthought, pushing accountability onto banks without ironclad enforcement.

The unions have argued that a top-down rollout without a dialogue with unions, consumers and civil society is a recipe for disaster. They have a valid point. Imagine public sector banks, already groaning under non-performing assets (NPAs), saddled with AI systems that hallucinate decisions or drift into bias, leading to wrongful loan denials to small borrowers.

Who foots the bill? The officers on the frontlines, scapegoated for vendor flaws, or the customers left in the lurch without a “right to human review”?

These concerns are not fear-mongering but a lesson in caution that Indian policymakers must draw from across the world while using AI. For instance, in the US, there are concerns over AI misuse that erodes jobs, demanding oversight and retraining. Australia mandates consultation before AI deployment to protect workers.

In this backdrop, India, with its public banks serving as lifelines for the underserved, can't afford to lag. All India Bank Officers' Confederation  demands — a national council for stakeholder talks, no forced redundancies, funded upskilling and mandatory disclosures for AI interactions — aren't anti-progress; they're pro-sanity.

Without these, private banks with deeper pockets can monopolise AI gains, eroding market share for PSBs and widening the chasm between urban elites and rural masses. Hence, the RBI's call for indigenous models and low-risk compliance tolerance is fine but it rings hollow if not backed by enforceable protections against model drift or data misuse.

The regulator must take a firm stand: encourage AI but also prioritise risks with the same zeal. In other words, innovation without accountability is reckless.

(Banking Central is a weekly column that keeps a close watch on and connects the dots regarding the sector's most important events for readers.)