Fake texts and screenshots in family court: what recent cases show, and what to keep

Last checked 12 November 2026. General information, not legal advice. TimeBinder does not detect fabricated messages or AI output, and a dated record is not proof that a message was sent or who wrote it.

What recent reports show

• A prescription photo and false testimony (Kentucky). In Adams v. Anderson, a divorce case in which one parent sought to change time-sharing, a parent testified about an urgent-care visit and submitted a photo of a prescription bottle. The other parent said the clinic and pharmacy had no records and that the documents were fake, altered or AI-made. The Kentucky Court of Appeals’ opinion records that the parent admitted lying to the court, that the family court held her in criminal contempt and sentenced her to 45 days in jail, and that the Court of Appeals affirmed on 29 May 2026 in an opinion designated not to be published. The opinion does not find that AI was used: the AI explanation was the other parent’s allegation, and the admission was that the testimony and documents were false or altered.
• Altered message images (California). Above the Law reports that in Mendones v. Cushman & Wakefield (Cal. Super., 2025), a civil case rather than a family case, the judge found, mainly by visual inspection, that two videos, three doorbell-camera photos and images of text message conversations had been materially altered with AI, and ended the case as a terminating sanction.
• A fake text and an arrest (Florida; a criminal matter, not a family case). WBBH reported on 21 October 2026 that a Naples woman is suing the Collier County sheriff and an officer. She says she was arrested for allegedly breaching a no-contact order on the strength of a text her ex-boyfriend supplied, which she says she never sent. Her defence lawyer pulled phone records that, per the report, showed the text was never sent or received, and the charge was dropped. The ex-boyfriend denies creating it, according to a Florida family-law firm’s summary. The lawsuit allegations are unproven. The defence lawyer told WBBH he now sees fake digital evidence in cases a couple of times a month.

These are individual cases from secondary reports. We cannot say how common fabricated messages are, and a claim that something is fake is also sometimes used to dodge real evidence.

How courts decide whether a message is real

In US courts the starting point is authentication: the side offering a message must produce enough evidence for a judge to find it is what they say it is (for example, Rule 901 of the Federal Rules of Evidence and state equivalents such as Florida’s section 90.901). The bar is deliberately modest, so a convincing fake can clear it, and the other side then has tools to challenge it.

• The rest of the thread, not one cropped image, plus dates, times and the context of the conversation.
• The device and the accounts. Carrier or platform records can confirm whether a message was sent and received, as in the Florida case above.
• Witnesses who were part of the conversation, and details in the message that point to a particular sender.
• Metadata and forensic examination of the original file or device, where the stakes justify it.

Rules of evidence differ by state and country. The federal rulemakers are still studying whether a deepfake-specific rule is needed; our piece on deepfake evidence and the rules explains where that stands.

If a message matters to your case: what to keep

1. The full thread, on the original device or account. Do not delete the original after taking a screenshot, and do not clean up the phone. Screenshots show only what the person taking them chose to capture.
2. Screenshots and exports made the same day, showing the contact name or number, dates and times, and surrounding messages. Our guide to screenshots as evidence goes through the limits.
3. A short note of when and how you received the message, who else saw it, and what you did next.
4. Related records: call logs, emails, voicemails, bills showing the number, and anything else that corroborates.
5. Do not edit, crop, enhance or re-save anything you may later offer. Keep working copies separate from the files you preserved.
6. Keep the set together and logged. A chain-of-custody log is a tidy way to record who held what and when.

If you think the other side has fabricated something, tell your lawyer rather than confronting them or investigating on your own. The next step may be carrier or account records, a formal request, or a forensic examiner. If you are accused of faking something real, our piece on what to do if the other side says your evidence is a deepfake may help.

What a TimeBind does and does not show here

A TimeBind is a dated, checkable record that an exact file existed no later than a recorded time and still matches. Your first Pooled TimeBind is free; see how it works. For a family-law matter, that means the screenshots, exports and notes you saved can be shown to be unchanged since the day you saved them. You can make a SHA-256 fingerprint of any file with our hash checker; the file stays on your device.

It does not show that a message was really sent, who wrote it, or that a screenshot is accurate. If you timestamp a screenshot of a fake message, the timestamp is real and the message is still fake. The value is a clean, dated record of what you held and when, kept alongside the original thread and carrier records, which do the proving. See our legal timestamping page and the guide for lawyers.

Limits

This is general information drawn from news and law-firm summaries, not legal advice. Evidence and family-court rules differ by state and country. Speak to a qualified family lawyer where you are before relying on any of it.

Tools that can help

There is no affiliate link on this page: nothing here depends on a paid third-party tool. Our affiliate disclosure explains when we do use them.

Sources checked

• Gulf Coast News (WBBH), Naples woman sues Collier sheriff, says fake text message led to her arrest, updated 21 October 2026.
• Hoffman, Larin & Agnetti, P.A., Can Your Ex Claim Your Texts Are AI-Generated in a Florida Custody Case?, 6 October 2026 (a law-firm marketing page).
• Above the Law, AI Evidence: Blurring The Line Between Reality And Fiction, 30 July 2026 (Mendones and Adams).
• Kentucky Court of Appeals, Adams v. Anderson, No. 2025-CA-1309-MR, rendered 29 May 2026, not to be published (copy of the opinion; also on Leagle).

Keep a dated record of what you saved

Your first Pooled TimeBind is free, with no credit card required. After that, TimeBinds start from US$5, and an ASAP TimeBind is US$15. Click here to start, or see plans and pricing.