Why Litigation Finance Firms Are Now Demanding AI Disclosure Appendices in Funding Applications

The litigation finance industry built its credibility on rigorous, independent case assessment. Funders like Burford Capital, Harbour Litigation Funding, and Omni Bridgeway carved out dominant positions by being better at predicting case outcomes than the lawyers bringing them. That analytical edge is now under threat — not from bad lawyers, but from good AI that produces confidently wrong answers.

Which explains why, quietly but with increasing urgency, major funders have started requiring something new in funding packages: an AI disclosure appendix.

The Problem Funders Are Actually Trying to Solve

Let's be direct about what's driving this. Funders are not worried about AI in the abstract. They're worried about a specific failure mode: AI-inflated merit scores dressed up as independent analysis.

Here's the pattern. A plaintiffs' firm runs their case theory through a large language model — Harvey, CoCounsel, a custom GPT, whatever — and gets back a confident, well-structured merits memo. The memo cites cases, walks through damages calculations, identifies comparable settlements. It reads like a senior associate who did the work over a weekend. The problem is that LLMs are optimized to produce coherent, plausible output, not accurate output. They hallucinate citations, misstate holdings, and — critically — tend to mirror the framing fed to them. If you describe your case with optimistic assumptions, you will generally receive optimistic analysis back.

This is not hypothetical. Since 2023, courts have sanctioned lawyers in multiple jurisdictions for submitting AI-generated briefs containing fabricated citations — Mata v. Avianca (S.D.N.Y. 2023) being the most cited example. The litigation finance context is arguably more dangerous than the courtroom context because there's no judge reviewing the work product for accuracy. The funder's investment committee is reviewing it, and they're relying on outside counsel's representations about methodology.

What the Disclosure Appendices Actually Require

The specific requirements vary by funder, but the appendices now being circulated by several major institutional funders share a common structure. They typically ask for:

Tool identification. Which AI platforms were used, and in what capacity — case research, merits analysis, damages modeling, comparable verdict analysis, document review. The distinction matters. Using AI for document review in underlying discovery is categorically different from using it to generate the probabilistic win assessment submitted to the investment committee.

Human review certification. A named attorney must attest that every AI-generated analysis was independently reviewed against primary sources. This isn't checkbox compliance — funders are increasingly asking for the name and seniority of the reviewing attorney, because they want to know whether a first-year associate spot-checked a damages model or a seasoned litigator did.

Input disclosure. Some funders are now asking for disclosure of the prompts or queries used to generate AI-assisted analysis. This is the most contested requirement, and several firms are pushing back on privilege grounds. But funders have a reasonable counter: if the AI output is being submitted as the basis for a multi-million-dollar funding decision, the methodological inputs are material.

Limitations acknowledgment. A brief, specific statement of what the AI tools cannot do in the context of this case — particular jurisdictional complexities, novel legal theories, or damages models that depend on expert opinion the AI cannot replicate.

How Plaintiffs' Firms Are Adapting

The better plaintiffs' shops have moved fast. Several mid-size litigation boutiques and legal ops teams at large law firms have restructured their funding application workflows to treat the AI disclosure appendix as a first-order deliverable, not an afterthought.

What this looks like in practice: case assessment teams now maintain what some are calling an "AI audit trail" — timestamped records of which tools were used, what was queried, who reviewed the output, and what changes were made as a result of that review. In some firms, this is being logged in matter management systems. In others, it's a simple but consistently maintained Word document that travels with the funding application.

The more sophisticated adaptation is at the damages modeling level. AI-assisted damages calculations — using tools that pull comparable verdicts, apply inflation adjustments, and generate ranges — are now being explicitly separated from the strategic judgment layer in funding packages. The AI generates the range. A named expert or senior partner provides the methodology footnote explaining why the AI-generated range is or isn't appropriate for this specific case. That separation is exactly what funders want to see.

Legal ops teams deserve particular credit here. At the firms that have gotten ahead of this, it's typically legal operations professionals who built the documentation protocol, not the litigators who were initially resistant to additional process overhead.

The Downstream Implications Are Significant

The AI disclosure requirement is doing something that courtroom sanctions alone couldn't accomplish: it's forcing internal quality control processes at the origination stage. If you know a funder's investment committee will scrutinize your AI methodology, you build the methodology to withstand scrutiny.

This will likely have a sorting effect on the market. Firms that can demonstrate rigorous, documented AI governance in their funding applications will be better positioned for capital access — and funders will increasingly treat AI disclosure quality as a signal of overall practice quality. Sloppy appendices suggest sloppy case management.

There's also a longer-term implication for how funding agreements are structured. We're already seeing early language in term sheets addressing AI-assisted case management post-funding — specifically, provisions requiring disclosure if AI tools are used to make material strategic decisions after capital is deployed.

The Funder-Counsel Relationship Is Being Recalibrated

What's really happening here is that funders are reasserting their analytical independence at the exact moment AI threatens to erode it. The disclosure appendix is how they're doing it. Plaintiffs' counsel and legal ops professionals who treat this as a compliance burden are missing the point. Funders are right to demand it, courts will eventually follow, and the firms that build clean AI governance now will have a structural advantage when this practice becomes standard. It's already almost there.