Remember that two-week interregnum when Rohit Chopra remained at the CFPB after the presidential transition?
While many assumed he would be gone on Day One, Chopra stayed just long enough to finish one last piece of work. On his way out, his team published a guide titled “Strengthening State-Level Consumer Protections.”
It was, effectively, a parting gift: a strategic playbook for states to enforce consumer protection laws if federal supervision went dark. It was a blueprint for how to assemble a state-level CFPB.
And now? Construction is officially underway.
The Rise of the “Shadow Director”
Rohit Chopra is back in the regulatory spotlight, but this time, he isn’t operating from Washington, D.C. A few days ago, the Democratic Attorneys General Association (DAGA) tapped Chopra to lead a new consumer protection working group.
The roster includes Attorneys General from California, New York, Massachusetts, Maryland, Illinois, Pennsylvania, Michigan, and more. Functionally, Chopra has become the “shadow director” of a decentralized, multi-state regulatory perimeter.
We are no longer looking at one CFPB; we are looking at many. And they are moving fast.
The Blueprint in Action: New Jersey’s N.J.A.C. 13:16
For lenders wondering what this new landscape looks like, the signal is clear. On December 15, New Jersey became the latest state to operationalize the Chopra roadmap by finalizing N.J.A.C. 13:16, its new rules on Disparate Impact Discrimination.
While these rules cover employment and housing, their impact on lending and AI is profound.
This is not an isolated legislative event. It is a coordinated pivot. Just three days before New Jersey finalized these rules, a coalition of 21 state Attorneys General formally opposed federal efforts to remove disparate impact provisions from the Equal Credit Opportunity Act (ECOA).
New Jersey’s framework is the first major implementation of this new strategy, and it changes the math for compliance officers nationwide.
The New Compliance Standard: “Less Discriminatory Alternative”
Under the new New Jersey framework, the burden of proof has shifted aggressively. If a lending practice disproportionately harms a protected group, it is a violation unless the lender can prove two things:
- The practice is necessary to achieve a legitimate business interest.
- 2. There is no “less discriminatory alternative” (LDA) available.
That second requirement is where lenders get tripped up.
Regulators have explicitly flagged rigid credit score cutoffs (e.g., automatically denying anyone below 600) as high-risk compliance triggers. The argument is: an individual assessment of a borrower’s ability to pay could achieve the same risk management goals with less discriminatory impact.
If you are using a rigid cutoff, a regulator can now ask: “Why didn’t you use cash-flow underwriting? Why didn’t you look at deposit data? You had a less discriminatory alternative, and you chose not to use it.”
A Shift in Regulatory Gravity
The states are sharing data, experts, and market intelligence at the exact moment when more consumer financial decisions than ever are being automated by AI systems.
This is already showing up in enforcement. Massachusetts levied a $2.5M penalty in an AI underwriting case this summer, and both California and New York are rapidly ramping up similar probes.
In practice, this shifts the center of gravity from one national “cop on the beat” to dozens of state enforcers. Each enforcer comes to the table with their own priorities, local politics, and novel theories of AI liability.
As AI scales decision-making, states are scaling oversight.
For lenders, the operational reality is stark: your models, decisions, and AI-driven workflows may now be tested not once in Washington, but many times—in Trenton, Sacramento, Albany, Boston, Annapolis, and beyond.
Strategic Adjustments: What Lenders Should Do Now
The federal retreat from lending supervision is real. The enforcement vacuum is not.
To navigate this decentralized regulatory environment, lenders should act now. Here is a strategic framework for the coming months:
- Operationalize “Alternative” Searches: It is no longer enough to justify your current model. You must demonstrate that you actively evaluated alternative approaches. If a regulator identifies a less discriminatory alternative that you failed to consider (like alternative data sources), your defense is compromised.
2. Leverage Cash-Flow Data: Regulators increasingly view cash-flow underwriting as a viable, less discriminatory alternative to FICO-based cutoffs. Incorporating consumer-permissioned cash-flow data can provide the necessary financial picture to satisfy the LDA requirement.
3. Simulate consumer outcomes under different thresholds and operating conditions: Evaluate consumer outcomes across different score cutoffs, policy rules, and operating conditions—at both the model and strategy level. This includes testing for disparities not just within the model itself, but across the full customer journey, where underwriting, pricing, and policy decisions compound. Done correctly, these simulations generate the empirical evidence needed to support credit decisions and allow institutions to calibrate risk appetite with precision, before regulators do it for them.
4. Design for the Strictest Standard: Design your compliance program for the strictest state standard, not the federal minimum. New Jersey moved first, but with the Chopra coalition in place, the infrastructure for other states to follow is already built.
Are You Ready for Rohit 2.0?
The landscape has changed. State-level enforcement requires state-level visibility.
If you are currently monitoring fair lending only at the national level, you are flying blind in jurisdictions like New Jersey, Massachusetts, California, New York, Maryland, Colorado and others.
See your state-level risk: [Request a Demo: Monitor Fair Lending by State] See how our platform isolates performance by geography to catch disparities before the AGs do.
Operationalize Your LDA Strategy: [Download our LDA Handbook: How to Choose a Less Discriminatory Alternative] Get the definitive guide on how to identify, test, and document alternatives.
