Large language models can draft chronologies in minutes. They can also invent citations, dates, and privilege boundaries. For funders underwriting non-recourse capital, that gap is a budget and waiver problem—not a novelty.
Where hallucinations hurt litigation
- Chronologies — wrong timestamps break settlement corridors.
- Disclosure lists — phantom documents trigger costly rework and waiver debates.
- Issue lists — counsel spends days correcting machine output instead of arguing merits.
Controls we require before funding
- Human sign-off — qualified lawyer approves any AI output used in pleadings or disclosure.
- Source pinning — every asserted fact links to Bates or native file hash.
- Versioning — prompts, model IDs, and outputs stored in the data room.
- Red-team pass — second reviewer hunts for invented cases and quotes.
For claimants: AI can shrink first-pass review cost. It cannot replace counsel judgment on privilege, strategy, or settlement authority.
See our x402 screening demo for how we structure ingest → merits → counsel review.