[Home](index.html) Why now
Why now
# Evidence got cheap to fake. Fast.
This is not a hypothetical risk being marketed at you. Insurers report it, courts are ruling on it, and the federal rules are being amended in response to it.
In short
In short: generative AI has made post-loss images easier to question at the same time insurers and courts are confronting manipulated records. HomeCloneVault addresses that trust gap by creating and sealing the property record before a loss and before any dispute exists.
## Key terms
**Pre-loss record**
Documentation created before damage and before a related dispute exists.
**Provenance**
The documented origin and handling history of a record.
**Tamper-evident**
Designed so a later file change produces a hash mismatch that can be detected.
Quick answer
99% of insurers say they have encountered manipulated or AI-altered documentation and only 32% say they are very confident they could identify a deepfake (Verisk State of Insurance Fraud study, March 2026), and courts have already identified deepfake evidence submitted as authentic. Proposed Federal Rule of Evidence 707 addresses only evidence a party admits is AI-generated — leaving disputed fabrications, which is the gap a pre-loss blockchain-sealed record is designed to fill.
- [The data](#data)
- [What courts are seeing](#courts)
- [Proposed Rule 707](#rule707)
- [Why the window is now](#tailwind)
- [FAQ](#faq)
## The data
99 %
of insurers say they have encountered manipulated or AI-altered documentation
[Verisk State of Insurance Fraud study, March 2026](https://www.verisk.com/company/newsroom/ai-editing-tools-are-fueling-a-new-era-of-insurance-fraud-according-to-new-research-from-verisk/)
32 %
of insurers say they are very confident they could identify a deepfake
[Verisk State of Insurance Fraud study, March 2026](https://www.verisk.com/company/newsroom/ai-editing-tools-are-fueling-a-new-era-of-insurance-fraud-according-to-new-research-from-verisk/)
76 %
of insurers say manipulated media submissions have grown more sophisticated
[Verisk State of Insurance Fraud study, March 2026](https://www.verisk.com/company/newsroom/ai-editing-tools-are-fueling-a-new-era-of-insurance-fraud-according-to-new-research-from-verisk/)
$ 45 B
annual U.S. property & casualty insurance fraud
[Coalition Against Insurance Fraud, 2022](https://insurancefraud.org/wp-content/uploads/The-Impact-of-Insurance-Fraud-on-the-U.S.-Economy-Report-2022-8.26.2022.pdf)
## What courts are seeing
| Deepfake video in court | Mendones v. Cushman & Wakefield — California, 2025 |
| --- | --- |
| Fabricated text messages | Florida, 2025 — two days in jail; charges dropped after eight months |
| Hallucinated case law | Arch Insurance Co. v. A3 Development — S.D. Florida, October 2025 |
| First federal AI-evidence rule | Proposed FRE 707 — released for public comment August 2025 |
| The gap Rule 707 leaves | Disputed deepfakes the proponent does not admit are AI-generated |
### Mendones v. Cushman & Wakefield
California, 2025
Judge Victoria Kolakowski identified AI-generated deepfake video submitted as authentic witness testimony — one of the first documented instances of a deepfake submitted as purportedly real evidence in a U.S. court.
### AI-fabricated text messages
Florida, 2025
A Florida woman spent two days in jail after AI-generated text messages were submitted as evidence against her. Charges were dropped after eight months. Cited by the National Center for State Courts (February 2026) as a threat to public trust in courts.
### Arch Insurance Co. v. A3 Development
S.D. Florida, October 2025
Attorneys cited court rulings that did not exist — AI hallucinations presented as real case law. The judge granted a motion to strike; the Florida Supreme Court subsequently issued new rules on AI use in court filings.
### Proposed Federal Rule of Evidence 707
U.S. Judicial Conference, August 2025
The first federal rule proposal specifically addressing AI-generated evidence. Critics note it applies only to evidence the proponent acknowledges was AI-created — it cannot close the gap of disputed deepfakes. That gap is what a pre-loss sealed record fills.
“Now do I, as a judge, have to question a source of evidence that has traditionally been reliable? We're in a whole new frontier.”
— Judge Erica Yew, Santa Clara County Superior Court, California, to NBC News · as cited by the [National Center for State Courts](https://www.ncsc.org/resources-courts/ai-generated-evidence-threat-public-trust-courts), February 24, 2026
## Proposed Federal Rule 707 — and the gap it leaves
In August 2025 the [U.S. Judicial Conference](https://www.uscourts.gov/rules-policies) released proposed Rule 707 for public comment — the first federal rule specifically addressing AI-generated evidence in court proceedings.
Critics have noted its limit: it applies to evidence the proponent acknowledges was AI-created. It does nothing about deepfakes and falsified evidence whose authenticity is disputed — the category that actually threatens a property claim. The gap Rule 707 cannot close is exactly the gap a pre-loss sealed record fills, because that record's timestamp is not a matter of anyone's testimony.
Florida courts, meanwhile, have issued new requirements on AI use in filings following a 2025 federal insurance case in which attorneys cited court rulings that did not exist.
## Why the window is now
- Post-loss documentation is no longer self-evidently trustworthy. The only evidence that cannot have been fabricated in response to a loss is evidence that was sealed before the loss — before any incentive to fabricate existed.
- Courts are legislating in real time. The legal system is actively searching for evidence standards it can trust, and independently verifiable timestamps are one of the few things that survive that scrutiny.
- Insurers are adding AI-related exclusions. As the evidentiary landscape shifts, policyholders holding pre-loss records are in a materially different position than those holding a camera roll.
- The value compounds backwards. A record sealed in 2027 becomes more persuasive over time precisely because it predates the period when everything became questionable.
## Seal a baseline before the next season.
Founding clients are being scheduled now ahead of the January 2027 launch in Southwest Florida.
[Join the waitlist](contact.html)
## Frequently asked questions
### Is AI-fabricated evidence actually showing up in real cases?
Yes. In 2025 a California judge identified AI-generated deepfake video submitted as authentic witness testimony in Mendones v. Cushman & Wakefield. In a separate Florida matter, AI-generated text messages led to a woman spending two days in jail before charges were dropped. The National Center for State Courts cited that case in a February 2026 report on AI evidence and public trust in courts.
### Does proposed Federal Rule 707 solve this?
Not for the hardest case. Rule 707, released for public comment in August 2025, is the first federal rule proposal specifically addressing AI-generated evidence, but critics note it applies to evidence the proponent acknowledges was AI-created. It does not resolve disputed deepfakes whose authenticity is contested — which is precisely the gap a pre-loss sealed record addresses.
### Why does waiting make this worse?
A baseline is only useful if it predates the event you are documenting. Every month without one is a month in which no pre-loss record exists for that property. Records created before synthetic media became trivial to produce also carry more persuasive weight, not less, as time passes.
Page published July 29, 2026 · last revised September 14, 2026 . Statistics on this page carry their original source and publication date.