The Deepfake Docket Has Arrived

On April 7, 2026, a thirty-seven-year-old man from Columbus walked into the Southern District of Ohio and pleaded guilty to three federal counts. Cyberstalking. Producing obscene visual representations of child sexual abuse material. And publication of digital forgeries — the statutory term the Take It Down Act uses for deepfakes.

The U.S. Attorney's Office publicly stated that James Strahler II is believed to be the first person in the United States convicted under the Take It Down Act. See United States v. Strahler, S.D. Ohio No. 2:25 cr (plea entered Apr. 7, 2026).

Strahler had installed more than twenty-four AI platforms and more than one hundred web-based AI models on his phone. Between December 2024 and June 2025 he used them to generate sexually explicit images and videos of real people without consent. He targeted at least six adult women. He distributed fabricated pornographic video to their coworkers. He sent voicemails containing threats of rape, paired with specific home addresses. He then produced synthetic child sexual abuse material using the faces of real minors from his own community and posted more than seven hundred such images to a website dedicated to that material.

The first person convicted under the Take It Down Act is from Ohio.

Not California. Not New York. Ohio.

I want Ohio practitioners to sit with that sentence for a moment. The deepfake docket did not arrive from the coasts. It arrived in our district, in our time, involving our neighbors. Whatever you believed about synthetic media last quarter, it is now a matter of Ohio criminal practice, Ohio evidentiary practice, and, inevitably, Ohio domestic relations practice.

This is a field report for the lawyers who practice in the courts where this problem actually lands.

Evidence Used to Reflect Reality. Now It Requires Verification.

For generations, the profession operated on a settled premise. A photograph captured a moment. A recording preserved a voice. A video told a story. Jurors trusted what they saw. Judges relied on recordings to reconstruct events. Investigators built prosecutions around surveillance footage.

Generative artificial intelligence has dissolved that premise.

Today, convincing audio can be fabricated from seconds of recorded speech. Entire conversations can be generated. Video of events that never occurred can be produced in minutes. Messages that appear authentic can be created without any human ever sending them.

The question our evidentiary system was built to answer was, "Is this what the proponent claims it is?" Under Ohio Evid. R. 901(A), and its federal counterpart Fed. R. Evid. 901(a), authentication is a low bar. The proponent must produce evidence sufficient to support a finding that the item is what the proponent claims.

That standard was calibrated for a world where fabrication required skill, time, and equipment. It was not calibrated for a world where a laptop can manufacture reality.

The result is not a hypothetical reform debate. It is a live operational problem in criminal cases, civil cases, employment cases, and, as I will show, in domestic relations cases.

The Case Law Is Already Forming

Practitioners often assume deepfake litigation is tomorrow's problem. The reported decisions suggest otherwise.

Place Huang, Khalilian, Rittenhouse, Kohls, Broadrick, Kramer, Schmidt, Smith, and Schaffer on the same desk. A pattern resolves. Courts are adjudicating synthetic media questions under authentication rules written for Polaroids. The results are, predictably, inconsistent.

The Liar's Dividend

The most corrosive consequence of generative media is not fabricated evidence. It is deniable evidence.

Consider a defendant presented with authentic video footage of admitted conduct. The historical response was denial of the conduct. The modern response, available to every defendant and every civil litigant, is denial of the medium.

That recording is fake.

That video was generated.

That message was synthesized.

In prior decades, such a denial was implausible. Today it is plausible on its face. The courts have begun calling this phenomenon the liar's dividend, and it has already appeared in family courts. A survivor of domestic violence produces a smartphone video capturing abuse. The accused spouse responds, without expert support, that the video must be AI generated. A custody outcome, a protection order, a criminal charge — each now turns on an evidentiary fight the survivor did not anticipate and may not be able to afford.

The liar's dividend is not a future problem for Ohio domestic relations practitioners. It is the next objection you are going to hear in a full hearing on a Civil Protection Order under R.C. 3113.31. It is the next objection you are going to hear in a contested allocation of parental rights under R.C. 3109.04. And when the magistrate calls for authentication, the cost of proving what everyone in the room already believes is going to fall on the party with the most to lose and the least resources to prove it.

That cost is the real harm. It is a tax on truth, paid by victims, and collected by the architecture of doubt that generative media introduces into every case.

Domestic Relations Is Ground Zero

Criminal deepfake prosecutions will generate headlines. Family court is where deepfakes will generate the most sustained damage.

The reason is structural. Domestic relations practice depends on evidence that is already informal. Text messages. Voicemails. Phone videos. Screenshots. Social media posts. Recordings made in the kitchen, the driveway, the parking lot. This is exactly the category of evidence that generative AI can most easily fabricate and most plausibly contest.

Anticipate the following within the next twelve months in Ohio domestic relations courts:

1. An AI-generated voicemail offered as threatening language in support of a CPO petition, with the alleged maker denying the voice entirely.
2. A fabricated text message thread offered in a contested custody hearing, with the maker asserting the thread was synthesized to manufacture a narrative of unfitness.
3. An AI-enhanced photograph offered in a parenting time dispute, with the opposing party invoking Wisconsin v. Rittenhouse to demand expert testimony on the enhancement pipeline.
4. A fabricated bank statement offered in an Affidavit 1 dispute. English family courts are already prosecuting litigants who submitted altered financial disclosures. See Vispute v. Vispute (Eng. 2023).
5. A fabricated intimate image deployed to humiliate or coerce a spouse during or after dissolution. The Strahler case shows the upper bound of that behavior. The lower bound, equally damaging to a divorcing family, will not make the U.S. Attorney's press release.

Every one of these scenarios becomes a hearing. Every hearing becomes expense. Every expense becomes a barrier to access to justice.

The Federal Rulemaking Response

The Judicial Conference's Advisory Committee on Evidence Rules has been working the problem. Two proposals have emerged.

The practical point for Ohio lawyers is this. Federal rule amendment is slow. Ohio's Evid. R. 901 mirrors the federal rule. Any Ohio Supreme Court response will follow, not lead, the federal process. That means trial judges, magistrates, and most often domestic relations magistrates are going to be adjudicating deepfake objections under the existing rules, with no circuit or Supreme Court of Ohio authority to guide them, for the foreseeable future.

This is an operational reality. It is also a strategic opportunity for practitioners who prepare.

The State Landscape, with Ohio Arriving

When S.B. 163 passes, Ohio joins the forty-plus states that have already acted. The question for practitioners is not whether your state will regulate synthetic media. The question is whether your practice is prepared when it does.

The Practitioner's Framework

Waiting for courts to define standards is not strategy. Acting before the standard arrives is. Here is the operational framework I am teaching at LegalTek.ai and deploying across my own firm.

The Real Risk Is Not the Fake. It Is the Erosion of Trust.

Technology introduces new tools. Law introduces new rules. Deepfakes introduce something deeper. Uncertainty about reality itself.

When juries question video. When investigators doubt recordings. When survivors are told their authentic footage must be authenticated before it can be heard. The justice system faces a challenge our evidentiary canon was not written to answer: preserving trust in evidence, and therefore trust in judgment.

Without that trust, litigation becomes procedural chaos and, more importantly, an engine of advantage for the party with more resources to contest everything. With that trust — preserved through updated rules, trained practitioners, and thoughtful governance — the justice system adapts.

The attorneys who understand synthetic media will lead this adaptation. The ones who do not will react to it, often from the losing side of a motion they did not see coming.

Final Thought

Artificial intelligence has not just changed how content is created. It has changed how truth must be proven.

The Strahler conviction in the Southern District of Ohio is not the end of this story. It is the opening paragraph.

The next chapters will unfold in CPO hearings. In contested custody trials. In financial disclosure disputes. In workplace investigations. In criminal prosecutions where the recording is either the whole case or the whole defense. And, eventually, in appellate decisions that will define the contours of Ohio Evid. R. 901 for a generation of practitioners.

The deepfake docket has arrived.

The question is whether your practice has.

Recommended Follow-On LegalTek.ai Articles

1. Authenticating Reality: A Trial Lawyer's Framework for Deepfake Evidence Under Ohio Evid. R. 901
2. The Liar's Dividend: When Defendants Claim Real Evidence Is Fake
3. AI Platform Liability: The Next Wave of Product Liability Litigation
4. Synthetic Evidence in Ohio Domestic Relations Practice: CPOs, Custody, and Affidavit 1 Disputes
5. The COUNSEL Framework: Governance, Disclosure, and Competence in the Age of Generative Evidence

Appendix A: Authorities Cited