The Law Already Says a Denied Loan Needs a Reason. Nobody Wrote That Rule for a Denied Job.

US credit law has required lenders to give specific reasons for a denial since the 1970s. NYC's Local Law 144 is one of the first attempts to bring a version of that standard to automated hiring — and it only mandates an audit, not an explanation to the rejected candidate.

NYC's Local Law 144 requires a bias audit of automated hiring tools, but not an individual explanation to a rejected candidate.

A rejected job applicant gets a form email with no reason attached, generated by an automated screening system somewhere upstream. A rejected loan applicant, by contrast, is legally entitled under the Equal Credit Opportunity Act to a specific adverse-action notice — a concrete reason code, not a vague form letter — a protection that has existed in US law since the 1970s, built in direct response to a documented history of lending discrimination.

That asymmetry isn't an accident of technology. Credit got explicit explanation rights early because redlining and discriminatory lending were well-documented enough to force a legal response. Hiring, benefits determinations, and tenant screening never received an equivalent, even though automated systems now make comparably consequential decisions in all three areas.

New York City's Local Law 144, effective since July 2023, is one of the more concrete recent attempts to close part of that gap — it requires employers using automated employment decision tools to commission an independent bias audit and to notify candidates that such a tool is in use. What it does not require is more telling than what it does: an individual candidate still has no legal right under the law to learn the specific reason their own application was screened out.

The EU AI Act takes a broader swing at the same problem, classifying certain employment and credit-scoring systems as high-risk and imposing documentation and transparency obligations on the companies that deploy them — a wider regulatory net than NYC's audit-only approach, though one still being tested in practice as enforcement ramps up.

The people most affected by this gap are the ones with the least power to push back: entry-level candidates, benefits applicants, borrowers without the resources to contest a decision through legal channels. A decision without an explanation cannot be meaningfully appealed — it can only be accepted or abandoned, which breaks a basic civic expectation that has existed, in some form, for as long as bureaucracies have made consequential decisions about individuals.

Interpretability research is producing tools that can extract at least partial, plain-language reasons from otherwise opaque scoring systems, and some employers are building explanation layers in voluntarily, ahead of any legal requirement, because an unexplainable rejection is becoming a real reputational and recruiting risk as awareness of laws like Local Law 144 spreads.

The realistic near-term standard is not full mechanistic interpretability of every hiring model in production — it's audits like NYC's expanding, piece by piece, into something closer to what credit law already guarantees: not just notice that an automated system was used, but a specific, actionable reason a real person can actually contest.