When I first joined the Consumer Financial Protection Bureau, we were still building the agency from the ground up: its mission, its systems, its very identity. I remember standing in the vestibule while leaders described our charge in words that still echo in my mind: “We’re building a 21st-century regulator.” 

That phrase meant something to me. Maybe because I think structurally — in data, systems and relationships — I heard speed, efficiency, accuracy. I heard an agency designed not to react, but to respond in real time. A regulator capable of shortening the distance between consumer harm and restitution. 

Fourteen years later, that vision is again on the line.

The Trump administration has revived its effort to weaken and potentially defund the CFPB. A proposal under the One Big Beautiful Bill would slash the bureau’s funding authority by nearly half, while parallel lawsuits question whether President Donald Trump can remove CFPB employees en masse. Enforcement has slowed, new rulemakings have been paused and uncertainty hangs over the agency’s future. 

The debate is often cast in ideological terms: government size, market freedom, regulatory overreach. But beneath that rhetoric is a different question, one that cuts across politics:

What if technology could deliver the same, or even better, consumer protection at a fraction of the cost?

A tech opportunity   

When the CFPB opened its doors, much of its work was done manually. Complaints were read by attorneys, tagged by analysts and triaged by hand. That process eventually led to over $20 billion in consumer relief, but it also demanded thousands of work hours and hundreds of staff. 

Today, AI can perform much of that work in seconds. A single model can read complaint narratives, detect patterns and categorize them with near-human accuracy. With proper training, it can even identify emerging risks before they surface publicly. 

This isn’t hypothetical. Banks are already using similar systems internally to monitor their own data, to detect when consumer harm may be systemic or when a control has failed. What once took an entire regulatory division can now be simulated by an intelligent system working continuously, without fatigue or bias. 

So, if AI can replace some of the heavy machinery of supervision, should oversight cost what it once did? Or, conversely, could a leaner regulator actually become a smarter one? 

AI oversight 

Savita Shankar, of Rutgers University, posed a simple but provocative question: “Can a central bank chatbot help consumer protection?” It’s an appealing image: a friendly digital interface answering questions and guiding consumers. 

But chatbots are just the surface. The real transformation happens beneath it, in the systems that read, reason and react across oceans of financial data. 

AI now serves as the analytical core of modern finance. It can detect anomalies across millions of transactions in milliseconds, distinguishing the unusual from the unlawful. 

 It can map every regulatory requirement to the data trails companies already produce, effectively turning laws into logic. It can prioritize enforcement or risk mitigation by severity and systemic exposure. 

In this sense, AI isn’t just a tool of conversation, it’s the operating system of modern oversight. 

The most dramatic gains from AI aren’t happening in enforcement, but in prevention. 

For fraud detection, for instance, AI has become indispensable. It studies behavior across accounts, devices and geographies, flagging synthetic identities and unauthorized transactions before losses multiply. Each improvement saves banks millions in write-offs and investigative labor. But it also protects consumers by catching harm before it happens, the same outcome regulation aims for. 

The same is true across the enterprise. AI models are refining credit decisions, predicting delinquency and automating servicing. They’re optimizing back-office processes, cutting operational costs and making financial products more responsive to how consumers actually behave. 

The result is a kind of dual-benefit efficiency: What’s good for the bottom line is often good for fairness. 

Efficiency and ethics 

When AI identifies bias in underwriting, it protects both the institution’s reputation and the consumer’s rights. When it flags a fraudulent transaction in real time, it reduces losses and restores trust. When it analyzes complaints to uncover systemic issues, it saves legal exposure while improving customer experience. 

For the first time, efficiency and ethics are converging. 

This raises a paradox that the CFPB debate has largely missed: If AI can dramatically lower the cost of oversight, both within banks and within regulators, then the question shouldn’t be how much government we can afford, but how intelligently we deploy it. 

A regulator equipped with AI could supervise continuously rather than periodically, detect issues before they escalate and direct human attention only where it’s truly needed. The potential savings are immense, not just in dollars, but in consumer outcomes. 

Yet there’s a danger in using this logic to justify budget cuts prematurely. Defunding before transformation risks hollowing out the agency before it can modernize. You can’t replace people with algorithms you haven’t built yet. 

Smart regulation  

Transitioning to an AI-enabled regulator demands investment in technology, governance, data integrity and the human expertise to guide it. It’s not deregulation, it’s smart regulation. 

The phrase “21st-century regulator” wasn’t just rhetoric. It was a blueprint for how oversight could evolve: Faster, fairer and more transparent. AI brings that vision within reach, not by replacing regulators, but by enabling them to focus where judgment still matters. 

In practice, this means regulators using AI to monitor systemic risk in near-real time, banks deploying AI to self-diagnose compliance gaps before examiners arrive, and consumers benefiting from faster resolutions and fewer preventable harms. 

If built responsibly, this ecosystem could shrink the distance between regulation and reality, between law and the lived experience of fairness in the financial marketplace. 

For decades, we’ve treated fairness and efficiency as competing goals. But AI is collapsing that divide. The most efficient systems are increasingly the fairest, because they detect bias, error and harm as inefficiencies to be eliminated. 

The market has already adapted. Now the regulator must. 

If the CFPB and financial supervisors more broadly embrace AI not as a cost-cutting measure but as a mission amplifier, they can finally deliver what we set out to build in that vestibule years ago: A regulator capable of protecting consumers at the speed of the marketplace. 

Whether that happens under a shrinking budget or a reimagined one remains to be seen. But one thing is clear, AI isn’t just transforming the business of banking. It’s redefining the business of trust. 

Jim McCarthy serves as chairman for McCarthy Hatch, which provides data-driven insights for risk management. A founding member of the Consumer Financial Protection Bureau, he is a keynote speaker and fractional CRO/CCO in the financial services industry with more than three decades of experience.