First-party residential property insurance relies heavily on photographic evidence. When a homeowner experiences damage from a storm, fire or plumbing leak, the usual process is to file a claim, submit pictures of the damage and await inspection by an insurance adjuster. This pathway to collect evidence assumes that photos accurately capture real damage and that the home remains in its damaged state until the adjuster arrives. But advances in artificial intelligence (“A.I.”) now allow homeowners to create deepfake images that convincingly depict nonexistent or exaggerated damage. Both legal and industry experts worry that this new technology makes it harder to rely on evidence in the claims process.
The threat is no longer hypothetical. Industry leaders in A.I., such as Verisk, warn that “digital media fraud” is emerging in claims, where A.I. generated or altered images substitute for authentic documentation of loss.[i] Swift Currie, a well-established firm that specializes in property-insurance law, has also warned that A.I. tools can fabricate property damage photos with seemingly genuine metadata, which consists of hidden data files that mimic what would normally be expected from a genuine photograph.[ii] Even consumer guides from insurers like Aviva warn of fake property photos being submitted to support fraudulent claims.[iii] The harms have become so prevalent that insurance companies are adopting deepfake detection software to counter these risks.[iv] These are just a few examples illustrating how A.I. has already entered the insurance fraud landscape.
The danger is exacerbated in states like Florida who has a liberal amount of time in their statutory inspection timeline. Under Florida Statute section 627.70131, insurers must begin investigating a claim within seven days of receiving the proof of loss, and if a physical inspection is required, it must be completed within thirty days.[v] This thirty-day window creates a real opportunity for manipulation. A homeowner can benefit by submitting A.I.-generated photos of alleged damage, quickly repair or conceal the issues and then claim compensation based on the repair estimate before the property adjuster ever arrives. By contrast, insurers are at a disadvantage: by the time an adjuster inspects the property, it may appear intact, leaving only the manipulated photographs as “proof” of the claim.
Such timing could create a litigation nightmare. Florida courts traditionally authenticate photos through witness testimony or proof of the reliability of the process that produced the photograph.[vi] However, A.I. can manipulate metadata and a witness may be willing to lie and “authenticate” the manipulated image. The burden is now on the insurer to prove fabrication.[vii] Judges may require heightened authentication, such as original raw files, edit histories or forensic expert review.[viii] However, without new evidentiary safeguards, the homeowner’s deepfake photos may survive a motion in limine and reach a jury.
Florida’s spoliation doctrine provides further complexity to this issue. In Martino v. Wal-Mart Stores, Inc., the Florida Supreme Court held that a plaintiff cannot bring an “independent first-party spoliation claim,” meaning a separate lawsuit, against the same defendant who both caused the injury and allegedly destroyed the evidence.[ix] Instead, when a party to the litigation destroys or alters evidence, the remedies are limited to procedural tools such as discovery sanctions or adverse inferences.[x] How it relates to this matter is in terms of destroying or altering evidence. If a homeowner repairs damages before the insurer’s inspection, the insurer cannot file a spoliation action. Rather, the insurer may argue that the missing evidence should be presumed unfavorable to the homeowner. In other words, the insurer could contend that the unrepaired condition would have weakened the homeowner’s claim, as an unaltered inspection might have exposed discrepancies in the alleged loss or demonstrated that the damage was less extensive than represented. The problem becomes more acute if fabricated evidence, such as A.I.-generated photos, is introduced in place of the missing proof. In that scenario, the jury could be misled and insurers may face significant challenges rebutting synthetic evidence without resorting to costly forensic analysis.
The industry has responded by developing detection tools. Forensic image analysis can reveal anomalies in lighting, pixel structure or metadata inconsistencies that suggest manipulation.[xi] Reverse-image searches are also used to catch recycled damage photos already in circulation.[xii] Discovery strategies increasingly demand production of original files, device information, cloud backups and contractor repair records. These steps are costly but necessary to guard against fraudulent deepfakes in residential property claims.
Florida’s statutory framework now intersects directly with this risk. The thirty-day inspection deadline, once a safeguard against insurer-delay, has become an exploitable gap for dishonest claimants.[xiii] Legislators may need to consider reforms requiring preservation of damage until inspection or presumptions against claimants who conduct premature repairs. Courts could also adopt heightened authentication requirements for photos offered in first-party insurance litigation. Without such measures, the rise of A.I. threatens to erode confidence in the claims process itself.
Ultimately, the intersection of deepfake technology and first-party residential property insurance reveals a fragile system. Homeowners’ claims have always been fact-driven, but when facts can be manufactured with convincing precision, the law must adapt. Insurers must accelerate inspection schedules, invest in forensic tools and press courts for stricter evidentiary standards. To avoid misunderstandings, plaintiffs’ attorneys should advise clients against premature repairs to ensure transparency in evidence submission. In this new era, seeing is no longer believing and the survival of fair insurance litigation depends on recognizing that truth.
[i] See Emily Law, Breaking Down Digital Media Fraud for Claims in the AI Era, Verisk (Aug. 06, 2025), https://www.verisk.com/blog/breaking-down-digital-media-fraud-for-claims-in-the-ai-era/ [https://perma.cc/E69P-9MKG].
[ii] See Melissa Segel & Kayla McCallum, Detecting and Combating Insurance Deepfake Fraud, Swiftcurrie (Mar. 20, 2024), https://www.swiftcurrie.com/newsroom-publications-Segel-McCallum-Detecting-and-Combating-Insurance-Deepfake-Fraud [https://perma.cc/7V7D-M3E4].
[iii] See AI and Deepafake Scams, Aviva (Jul. 2024), https://www.aviva.co.uk/help-and-support/protect-yourself-from-fraud/knowledge-centre/deepfake-and-ai-scams/ [https://perma.cc/RY4Y-UPZ6].
[iv] See Nicos Vekiarides, Viewpoint: Deepfake Fraud is On the Rise. Here’s How Insurers Can Respond, Ins. J. (July 17, 2024), https://www.insurancejournal.com/news/national/2024/07/17/784226.htm [https://perma.cc/L3TS-AZDM].
[v] See Fla. Stat. § 627.70131(3)(a)–(b)(2025).
[vi] See City of Miami v. Kho, 290 So. 3d 942, 944–45 (Fla. 3d DCA 2019).
[vii] See Serrano v. Citizens Prop. Ins. Corp., 368 So. 3d 1064, 1065 (Fla. 3d DCA 2023).
[viii] See Natalie Runyon, Deepfakes on Trial: How Judges are Navigating AI Evidence Authentication, Thomson Reuters Inst. (May 8, 2025), https://www.thomsonreuters.com/en-us/posts/ai-in-courts/deepfakes-evidence-authentication/ [https://perma.cc/DY4Z-QR63].
[ix] See Martino v. Wal-Mart Stores, Inc., 908 So. 2d 342, 345 (Fla. 2005).
[x] Id. at 346.
[xi] See Hannewacker v. City of Jacksonville Beach, 419 So. 2d 308, 311 (Fla. 1982).
[xii] See Law, supra note i.
[xiii] See Fla. Stat. § 627.70131(3)(b)(2025).
