I’ve been wrestling with a question that feels deceptively simple but grows more complex the longer you sit with it:
Will AI make accountability in banking better—or worse?
At first glance, it feels like it should make things better. AI is fast, consistent, and unfazed by fatigue. It doesn’t blink at volume, it doesn’t get bored with rules. In theory, AI can help make smarter credit decisions, catch fraud faster, stay on top of compliance, and flag what humans might miss.
But if you peel back the layers, accountability isn’t getting clearer—it’s just getting more... complicated.
We already operate within a dense thicket of checks and balances: internal controls, second-line functions, external auditors, and then the regulators. But despite all of that, when something breaks, responsibility is still hard to pin down. The response usually lands somewhere between “not my jurisdiction” and “run faster than the blame.”
And now we’re adding AI into the mix—a system even its creators sometimes struggle to fully explain.
Let’s not pretend that saying “a human is ultimately responsible” settles the matter. Most people working with AI systems don’t truly understand how decisions are made—because often, no one does. Not even the AI itself. It just... works. Until it doesn’t.
The Coach Problem
It’s like being a football coach.
You can own the outcome—but you didn’t make the pass, miss the goal, or worse, elbow the opponent in the face and get a red card.
That’s the complexity of accountability in large systems. You didn’t do it, but it was done by someone, and that could be complex.
With AI, we’re entering that space. The question is: Should accountability be voluntary or enforced? Moral or legal?
Zoom out, and yes, the Board of Directors holds ultimate responsibility. But let’s be real—most boards weren’t built for this new era. I’ve seen it up close: experienced board members suggesting solutions from the 1980s with full conviction, only for the operational team to quietly explain (after a few rounds of ATRs) that the world has moved on—and so have customer expectations.
A Quiet Shift in Lending Accountability
We’ve already been through one wave of this.
When digital lending took off, the game was all about quick sanctions. Approvals in minutes to over Rs 2 trillion. A race to “yes.”
But that phase matured. The industry slowly realized: Sanction is only the beginning.
Real accountability lies in what happens after the money leaves your system—monitoring repayment behaviour, anticipating stress, stepping in early, collecting with empathy and discipline.
The old model placed all the weight on underwriting. But the evolved paradigm balances credit approval and post-disbursal monitoring—and this is where accountability today still falters. That gap is not just operational—it’s structural.
The Hired Gun Fallacy
The real risk now is that AI becomes the consultant you blame.
“Oh, the algorithm said so.”
“The model picked that borrower.”
“That outcome? It was the system logic.”
The finger’s on the trigger, but no one knows who pulled it—or why.
The irony is, AI is already outperforming humans in many specific tasks—detecting fraud, segmenting customers, predicting delinquency, optimizing pricing. These aren’t marginal improvements. These are 1000x step-changes.
And so the paradox is: the only way forward may be more AI—not less.
AI to monitor AI.
AI to test controls.
AI to predict failure before it escalates.
The Era to Build (and Redefine Accountability)
Everything I see around—the credit engines, fraud stacks, compliance workflows, marketing tools—feels like a remix of 2010s thinking. Marginal upgrades. Nothing fundamental.
But that’s why this moment is so exciting.
This is not just the era to build new platforms.
This is the era to build new accountability architectures.
Because we are not going back to a world where humans alone can supervise complex systems. Nor can we hide behind AI as a hired gun. The only viable future is one where AI becomes part of the control mechanism—and part of the accountability loop.
That’s the hard part. But also the opportunity.
The accountability game is changing. Let’s not sit on the sidelines.
