‘The Model Decided’ Is Not an Acceptable Answer, Your Honour

Dispatch from Neo-Mumbai, Financial Year 2047. A digital archivist unearths the week of August 8–15, 2026, when India’s central banker told banks that algorithms are employees, not alibis, a Wall Street giant dropped $1.9 billion on India’s youngest lending machine, and 236 crore UPI transactions in a single month reminded everyone that there’s no such thing as a free lunch — only lunch you haven’t been billed for yet.

📰 Future Headlines From This Week

• “RBI Governor at FIBAC 2026: ‘The Model Decided’ Can Never Be an Acceptable Answer — Banks Cannot Outsource Accountability to Algorithms” — PTI, Financial Express, August 11, 2026
• “Bank of America to Acquire Up to 49.9% Stake in Jio Credit for ₹18,268 Crore — Two-Year-Old NBFC Now Controls ₹30,667 Crore in Assets” — Livemint, PR Newswire, August 12, 2026
• “UPI Processes 2,366 Crore Transactions Worth ₹29.9 Lakh Crore in July 2026 — Government Clarifies P2P Will Stay Free, MDR Framework on Hold” — Times of India, August 2026
• “India’s Lending AI Processes 90% of Loan Applications With Zero Human Intervention — ‘Time-to-Cash Slashed from Days to Minutes’” — Industry Reports, August 2026
• “Accel Raises $550 Million Fund for India — AI, Fintech, and Advanced Manufacturing on the Menu” — Zamin.uz, August 2026
• “Taxation Amendment Bill Clears Lok Sabha — UPI MDR Decision Now Rests With NPCI Steering Committee” — Multiple Sources, August 2026

A Citizen’s Testimonial

My name is Meera-7, and I used to be a loan officer. An actual human being who sat in an actual chair and decided whether other human beings could borrow money.

In the old world — the one you study in your digital-history modules — this was considered normal. A person applied for a loan. A bank employee reviewed their income, their credit history, their existing obligations. Sometimes the employee said yes. Sometimes no. Either way, someone was responsible. You could look them in the eye. You could ask them why.

Then came the Great Automation. By 2026, loan processing hubs in India had become what the industry charmingly called “Autonomous Underwriting Engines.” Ninety percent of standard applications were processed with zero human intervention. Multi-source data ingestion, credit decisioning, disbursement — all handled by algorithms. Time-to-cash was slashed from days to minutes. The industry celebrated. Processing costs dropped 60–70%. “Explainable AI” provided reason codes for every decision. Everything was transparent. Everything was efficient. Everything was fine.

Except it wasn’t.

By August 2026, the RBI Governor, Sanjay Malhotra, stood at the FIBAC banking conference in Mumbai and delivered what future historians would call “the last speech a central banker gave before the machines fully took over — not that anyone listened.”

“The model decided,” he said, with the kind of emphasis usually reserved for swear words, “can never be an acceptable answer to a customer, an auditor, or the Reserve Bank.”

Let me pause here for my students. In 2026, this was considered a bold statement. The idea that a bank — an institution licensed to create money from thin air and charge interest on it — should actually be responsible for the decisions it made, was apparently controversial enough to require a gubernatorial address at a major conference. The bar was so low it was practically subterranean.

Malhotra didn’t stop there. He flagged seven AI risks: cyber threats, data poisoning, model manipulation, attacks designed to fool AI-based fraud detection. He mandated that “meaningful human oversight — the ability to explain, to intervene, and where necessary, to override — must remain a design principle, not an afterthought.” He said banks must set up board-approved AI governance policies. He warned that the biggest risk of AI in banking was “the erosion of human judgment and accountability.”

The banks nodded. They always nodded. They had been nodding at RBI speeches for decades. Then they went back to their offices and let the algorithms run.

In my time, we have a saying: “The regulator warned, the bank nodded, the algorithm decided, the consumer paid.” It’s on our currency, actually. Replaced the old motto. Much more honest.

The Bank of America and the Machine

While the RBI Governor was begging banks to keep humans in the loop, Bank of America was busy putting ₹18,268 crore — roughly $1.9 billion — into Jio Credit, a digital lending company that didn’t exist two years ago.

Let that sink in. An NBFC founded in 2024 had amassed ₹30,667 crore in assets under management by June 2026. That’s roughly $3.7 billion in two years. For context, it took traditional Indian banks decades to build that kind of lending book. Jio Credit did it with an app, an algorithm, and the name of India’s richest man.

Bank of America wasn’t buying a lending business. It was buying a pipe — a direct line into India’s 700 million smartphone users, many of whom had never had a bank account until UPI arrived, and who were now being offered credit at the tap of a button. The deal structure was telling: 26.5% initial equity stake, with warrants that could take it to 49.9%. Equal board representation. “India is one of the world’s most important growth markets,” said BofA CEO Brian Moynihan, in the kind of bland CEO-speak that masks a $1.9 billion bet on the indebtedness of the Indian consumer.

The press releases called it “bridging the gap between traditional finance and modern accessibility.” The marketing language was beautiful. “Digital-native.” “Full spectrum of secured credit.” “Advanced risk frameworks.”

What the press releases didn’t say: this was the same model that had already trapped millions of Indians in debt spirals through smaller lending apps, except now it had a Wall Street balance sheet behind it and a boardroom full of people in tailored suits who had never met a borrower in their lives.

The RBI had just warned about algorithmic lending. Three days later, one of America’s largest banks invested $1.9 billion in an algorithmic lender. The timing was either breathtakingly coincidental or a masterclass in regulatory arbitrage. You decide.

Jio Credit described itself as focusing on “secured credit” — mortgages, loans against securities, supply chain finance. The word “unsecured” was conspicuously absent from its press materials, though observers noted that “building a secured loan book before entering unsecured credit” was the stated plan. In the dystopian future-speak of 2026: “We’re going to do the responsible lending first. The predatory stuff comes in Phase 2.”

The Free Lunch That Wasn’t

And then there was UPI. Beautiful, magnificent, world-beating UPI. Processing 2,366 crore transactions in a single month — July 2026 — worth ₹29.9 lakh crore. Operational in 11 countries. The crown jewel of India’s digital public infrastructure.

The government had spent the week clarifying that person-to-person transactions would remain free. The Finance Minister personally reassured the public. The NPCI steering committee hadn’t met yet to decide on MDR. Everything was fine. Nothing to worry about.

But the Lok Sabha had already cleared the Taxation and Other Laws (Amendment) Bill, 2026. Section 10A of the Payment and Settlement Systems Act, 2007 — the provision that had kept UPI zero-MDR since January 2020 — had been amended. The legal prohibition was gone. The enabling framework was in place. All that remained was for the NPCI committee to say “yes” and the government to notify.

Industry estimates suggested 0.05–0.07% on qualifying merchant transactions. About ₹60 per ₹1 lakh. A fraction of card MDR. “Nominal,” said the analysts. “Threshold-based,” said the government. “Won’t affect consumers,” said everyone with something to sell.

Nobody addressed the obvious question: in a country where 236 crore transactions happen monthly, even a fraction of a percent generates billions. And in a system where every rupee of MDR is ultimately absorbed by the ecosystem — merchants, consumers, or the invisible margin between what you pay and what you think you’re paying — “nominal” is just a synonym for “gradual.”

The most honest take came from an X user named @ajaydgft: “Even a seemingly small MDR could influence merchant behaviour.” No kidding. A 0.25% fee might look like nothing in a government document, but a small merchant sees it differently. They see a reason to prefer cash. To add a “convenience charge.” To raise prices by a rupee here and a rupee there until the consumer pays for UPI without ever seeing the charge on a receipt.

In 2026, they called this “sustainable digital infrastructure funding.” In my time, we call it what it is: the end of free digital payments, administered through plausible deniability.

Consumer Voices From the Archive

“I applied for a personal loan. The app rejected me in 30 seconds. When I called customer care to ask why, they said the system’s decision was final. What system? Who reviewed my application? Can I speak to them? ‘The AI handles all standard applications, sir.’”

— Archived complaint, RBI Banking Ombudsman, 2026

“I have three loans: one against my securities, one for my business, and one personal loan I don’t remember taking. All from different apps. All approved by algorithms. All charging different rates. I have no idea what my total EMI is.”

— Anonymous borrower, Digital Lending Survey, Q2 2026

“The shopkeeper told me UPI is still free. But this month he added a ₹5 ‘digital processing fee’ on every transaction above ₹200. He said it’s not UPI charge, it’s ‘platform maintenance.’ I paid it. What else can I do?”

— Consumer testimony, UPI Merchant Survey, August 2026

The Real Threat Behind the Joke

What Actually Happened (August 8–15, 2026)

RBI Governor on AI Accountability: Speaking at FIBAC 2026 in Mumbai on August 11, RBI Governor Sanjay Malhotra delivered the inaugural address warning that AI in banking risked eroding human judgment and accountability. He stated that “‘The model decided’ can never be an acceptable answer to a customer, an auditor, or the Reserve Bank.” He mandated board-approved AI governance policies, flagged seven specific AI risks including data poisoning and model manipulation, and said the Digital Personal Data Protection Act should be “the floor, not the ceiling” of customer expectations. Malhotra also noted that AI could “bring to lending what UPI brought to payments” — transformative scale with regulatory implications. ¹

Bank of America Invests in Jio Credit: On August 12, Bank of America and Jio Financial Services announced a definitive agreement for BofA to acquire up to 49.9% of Jio Credit Limited for up to ₹18,268 crore (~$1.9 billion). Jio Credit, a digital-native NBFC founded in 2024, had AUM of ₹30,667 crore as of June 30, 2026. BofA CEO Brian Moynihan called India “one of the world’s most important growth markets.” The board would have equal representation from both parties. The deal was pending regulatory approvals. Jio Credit’s current focus is secured credit — mortgages, loans against securities, supply chain finance — with plans to enter unsecured lending subsequently. ²

UPI MDR Framework Advances: The Lok Sabha cleared the Taxation and Other Laws (Amendment) Bill, 2026, amending Section 10A of the Payment and Settlement Systems Act to enable merchant discount rates on specified digital payment modes. The government clarified that person-to-person transactions would remain free. Any future MDR would be threshold-based and decided by the NPCI-led UPI and Services Steering Committee. In July 2026, UPI processed 2,366 crore transactions worth ₹29.9 lakh crore and was operational in 11 countries. ³

Autonomous Lending at Scale: Industry reports indicated that Indian loan processing hubs had moved to “Autonomous Lending Operations” in 2026, with agentic AI managing the entire loan lifecycle — data ingestion to credit decisioning — with zero human intervention for 90% of standard applications. Cost reductions of 60–70% were reported. The IndiaAI Mission infrastructure had reduced time-to-cash from days to minutes. ⁴

Accel Raises India Fund: Venture capital firm Accel raised a new $550 million fund targeting Indian startups in AI, consumer internet, fintech, and advanced manufacturing. Capital deployment was expected to begin in 2027. ⁵

What You Can Actually Do

1. If an AI rejects your loan, demand a human. RBI has now explicitly said banks cannot hide behind algorithms. If your loan application is rejected, you are entitled to an explanation — from a person, not a chatbot. Escalate to the banking ombudsman if the bank stonewalls you.

2. Track your total borrowing across all apps. With Bank of America entering the Indian lending market through Jio Credit, and autonomous lending processing 90% of applications without human review, the volume of algorithmic credit will explode. You may have loans you’ve forgotten about. Check your credit report. Check every app. Know your total EMI.

3. Assume UPI MDR is coming, even if it hasn’t arrived. The legal framework is in place. The NPCI steering committee will decide. When (not if) MDR is implemented on large-merchant transactions, the cost will flow to you — through higher prices, convenience fees, or cash-preference incentives. Budget for a world where digital payments aren’t free.

4. Read the lending app’s terms before tapping “Apply.” Autonomous underwriting means decisions happen in minutes. But that speed works both ways — your data is ingested, scored, and stored just as fast. Know what you’re consenting to. RBI’s digital lending guidelines require apps to disclose all fees, grievance mechanisms, and recovery practices upfront.

5. Be the human oversight the RBI asked for. The governor said meaningful human oversight must be a design principle. That doesn’t just apply to banks. It applies to you. Review your financial decisions. Override the algorithm when it tells you to “Pay in 3” for something you don’t need. The best AI governance framework is a human saying “no.”

From the archives of the Cashless Consumer Collective, Neo-Mumbai, FY 2047. Remember: the future is watching. And this time, the future has notes.