For decades, an unrepresented borrower fighting foreclosure meant a familiar playbook: “produce the note,” a handwritten answer, a boilerplate motion, and an easy path to summary judgment. That playbook is disappearing.
Generative AI tools like ChatGPT and Claude are now free or nearly free, and borrowers are using them as de facto legal drafting assistants. The result is a new generation of pro se filings that look — and often read — like they came from counsel.
What’s Actually Changing
The shift from “before AI” to “after AI” is stark:
- Generic boilerplate arguments have given way to Regulation X and dual-tracking arguments, loss mitigation procedural attacks, and chain-of-title challenges.
- Sloppy handwriting has given way to polished grammar, organized headings, and structured legal arguments.
- Vague complaints have given way to specific statutory citations — RESPA, FDCPA, CFPB, and HUD allegations that force a substantive response.
This isn’t an isolated trend. Attorneys nationally are reporting multiple cases with AI-generated filings in a matter of minutes — a pattern some are calling “AI slop,” where a striking share of cited case law simply doesn’t exist. But even the frivolous filings aren’t free: they delay sales, trigger continuances, and drive up briefing costs.
Why It’s Dangerous — Even When It’s Wrong
What makes AI-assisted filings a genuine operational risk isn’t that they’re always right. It’s that they’re often plausible. They use correct legal terminology, target real operational weak points like escrow errors and notice defects, and are long and organized enough to consume significant attorney review time — regardless of merit.
The key for servicers and their counsel is learning to separate the two categories quickly:
- Real exposure: Regulation X and dual-tracking violations, specific CFPB notice-timing defects, payment application errors, genuine assignment timeline gaps.
- AI noise: hallucinated case citations, out-of-jurisdiction authority, recycled “produce the note” arguments, and sovereign-citizen-adjacent theories dressed up in proper legal format.
Where the Risk Concentrates
AI use is not evenly distributed. AI-assisted filings are most active in judicial foreclosure states with long timelines, active foreclosure-defense bars, and mediation or settlement-conference programs — think Florida, New York, New Jersey, Illinois, Ohio, and Pennsylvania. High-volume bankruptcy districts amplify the effect further, often mattering more than a state’s foreclosure procedure itself. Even nonjudicial states like California and Texas are seeing growth through bankruptcy filings and post-sale litigation.
The common thread across every hotspot: more court touchpoints and more consumer-protection statutes simply give AI more material to work with.
Spotting an AI-Assisted Filing
A few reliable tells: overly formal tone inconsistent with the borrower’s evident knowledge at the hearing, citations to case law that doesn’t exist, “kitchen sink” statutory theories thrown in simultaneously, and borrowers who can’t explain their own arguments in court. Judges are catching on — some courts now require citation certification before a hearing, and sanctions have already been imposed for fabricated authority in both state and bankruptcy courts.
How Servicers and Their Counsel Can Respond
- Tighten the file before litigation starts. AI-assisted borrowers are good at finding weak files — endorsements, assignment recordings, notice documentation, and escrow reconciliations all need to be clean before the first pleading is filed.
- Standardize evidentiary packages. Boarding records, assignment timelines, and Regulation X compliance notes should be built once and ready to produce on demand, not assembled reactively.
- Respond early and firmly. Hallucinated citations deserve prompt Rule 11 or sanctions motions — courts are increasingly receptive to that argument.
- Train attorneys on the substance. Associates who can quickly separate a genuine CFPB or Regulation X issue from AI-generated noise will save real time and money.
- Fight AI with AI. Servicers and firms can use the same technology internally — to verify citations, flag likely AI-generated arguments, and draft first-response templates faster.
The Bottom Line
AI isn’t going to turn most pro se borrowers into skilled foreclosure defense attorneys. But it is permanently raising the baseline sophistication of what shows up on the docket. The cost to servicers isn’t usually a loss on the merits — it’s delay, motion practice, and the resources spent sorting signal from noise.
The firms and servicers who adapt fastest — with clean files, standardized evidence, trained attorneys, and their own technology countermeasures — will be the ones who keep this trend from becoming a real drag on foreclosure timelines and cost.
Back to all insights