☎ Call Today
Criminal Defense • Frisco, Texas
Serving 9 DFW Counties — Collin • Dallas • Denton • Tarrant • Rockwall • Kaufman • Ellis • Johnson • Hunt — Available 24/7
The L and L Law Group team at our Frisco, Texas office — co-founding partners Reggie London and Njeri London with staff
Our Frisco officeEst. 2011
The L and L Law Group team·Frisco, Texas
Sex Crimes · AI-Generated CSAM

Texas and federal AI-generated CSAM defense

And federal AI-generated CSAM cases in Texas often turn on details — the stop, the paperwork, the deadlines, the forum. What happens in the first weeks after arrest often matters as much as what happens at trial. L and L Law Group defends these cases in courts across North Texas, including Collin and Denton Counties.

AI-generated child sexual abuse material (CSAM) — synthetic imagery produced by generative-AI tools depicting fictional minors — is the most rapidly evolving area of state and federal sex-offense prosecution. Charges proceed concurrently under Texas Penal Code § 43.26(a-1) (expanded by HB 2700 in 2023 to reach synthetic imagery) and under federal 18 U.S.C. §§ 2252A and 1466A, the latter of which survives Ashcroft v. Free Speech Coalition by anchoring liability in obscenity doctrine rather than the prior virtual-pornography ban. Federal exposure includes a 5-year mandatory minimum for receipt or distribution and a 15-year mandatory minimum on a prior conviction — making early-phase strategy, suppression of digital evidence, and First Amendment as-applied analysis the highest-leverage moves in any DFW federal or state case.

and federal AI-generated CSAM: Texas punishment ranges at a glance
Offense levelConfinementMax finePenal Code
Class A misdemeanorUp to 1 year, county jail$4,000§12.21
Third-degree felony2 – 10 years, TDCJ$10,000§12.34
Second-degree felony2 – 20 years, TDCJ$10,000§12.33
First-degree felony5 – 99 years or life, TDCJ$10,000§12.32

Ranges per Tex. Penal Code ch. 12. Enhancements, deadly-weapon findings, and prior convictions can raise the applicable range; some offenses carry their own special ranges.

16 min read 3,700 words Reviewed May 17, 2026 By Reggie London
Direct Answer

AI-generated child sexual abuse material is prosecuted concurrently under Texas Penal Code § 43.26(a-1) (added by HB 2700 in 2023 to reach synthetic imagery) and under federal 18 U.S.C. §§ 2252A and 1466A. Federal § 2252A operates through § 2256(8)(B), which defines "child pornography" to include any computer-generated image "that is, or is indistinguishable from, that of a minor engaging in sexually explicit conduct" — receipt or distribution carries a 5-year mandatory minimum and 15 years on a prior; production under § 2251 carries a 15-year mandatory minimum. Federal § 1466A is the obscenity-based companion that reaches drawn, cartoon, and AI imagery regardless of whether any real child was depicted, anchored in Miller v. California obscenity doctrine to bypass the First Amendment limitation that Ashcroft v. Free Speech Coalition, 535 U.S. 234 (2002), imposed on the prior virtual-pornography ban. Texas § 43.26 grades 3rd-degree felony (default) to 2nd-degree (6+ items) to 1st-degree (50+ items or victim under 14) under § 43.26(d), with lifetime ch. 62 registration on conviction. Defense work centers on First Amendment as-applied challenges (the surviving Ashcroft question for fully synthetic imagery), Fourth Amendment suppression under Riley, Carpenter, and Pope v. State, mens-rea contests on knowing possession, Daubert challenges to forensic AI-detection expert testimony, and charge-bargaining between § 2252A and § 1466A. Local DFW practice involves HSI Cyber Crimes Center, North Texas ICAC, FBI Dallas Cyber Squad, and concurrent prosecution in NDTX (Dallas Division) and EDTX (Sherman/Plano divisions).

Free case review
Key Takeaways
  • Federal mandatory minimums — § 2252A(b)(1): 5 years for receipt/distribution, 15 years on prior; § 2251: 15 years for production.
  • Texas § 43.26(a-1) — HB 2700 (2023) expanded the statute to reach AI-generated/synthetic imagery without requiring a real child.
  • Two federal pathways — § 2252A via § 2256(8)(B) "virtually indistinguishable" test; § 1466A via Miller obscenity doctrine.
  • Ashcroft v. Free Speech Coalition — 535 U.S. 234 (2002), is the surviving First Amendment as-applied issue for fully synthetic imagery.
  • Lifetime registration — Texas ch. 62 and federal SORNA both apply; collateral consequences often outweigh the custodial sentence.
Quick Case Review · 24/7

Get a free review

Direct to attorney — no call center. Most clients hear back within an hour.

By submitting, you agree to our Privacy Policy. No attorney-client relationship is formed until a written engagement is signed.

Texas Bar
Licensed since 2004
TXND · TXED
Federal Court Admitted
4.8 ★
Google Reviewed
9 DFW
Counties Served
24/7
Direct-to-Attorney Line
40+
Years Combined
Texas Bar Licensed TXND & TXED Federal 24/7 Jail Release Se Habla Español
Texas Legal Context

What the statute actually requires

Analytical framework AI-generated CSAM is the most rapidly evolving area of state and federal sex-offense prosecution in 2024-2026. Federal liability arises through two distinct statutory pathways: 18 U.S.C. § 2252A (via the "virtually indistinguishable" test in § 2256(8)(B)) and 18 U.S.C. § 1466A (via obscenity doctrine under Miller v. California). Texas § 43.26(a-1), added by HB 2700 in 2023, runs parallel to the federal framework and reaches synthetic imagery on the same penalty schedule as actual-child material. The constitutional landscape is shaped by Ashcroft v. Free Speech Coalition, 535 U.S. 234 (2002), which struck down the original virtual-pornography ban — Congress responded with the PROTECT Act of 2003 and the Adam Walsh framework, but the as-applied First Amendment question for fully synthetic AI imagery remains live.
5 Texas-specific insights
  1. Federal mandatory minimums make federal court the dispositive forum. § 2252A(b)(1) imposes a 5-year mandatory minimum for receipt or distribution, escalating to 15 years on a prior chapter-110 conviction. § 2251 production carries a 15-year mandatory minimum, escalating to 25 years on one prior and 35 years on two priors. Texas § 43.26 has no mandatory minimum and tops out at 1st-degree felony exposure (5-99 or life), but the structural absence of a federal-style mandatory minimum often makes the state-court forum the more favorable one even at the 1st-degree level. Defense work in the early investigative phase often focuses on positioning the case for state-court rather than federal-court charging.
  2. § 2256(8)(B) "virtually indistinguishable" test is the federal gateway. 18 U.S.C. § 2256(8)(B) defines "child pornography" to include any "computer-generated image that is, or is indistinguishable from, that of a minor engaging in sexually explicit conduct." The statutory definition of "indistinguishable" in § 2256(11) places the perception in the eye of an "ordinary person viewing the depiction." This is the textual hook by which AI-generated imagery enters § 2252A. United States v. Spencer, 113 F.4th 1129 (8th Cir. 2024), is the leading recent appellate authority on the application of this framework to AI-generated imagery.
  3. § 1466A bypasses Ashcroft by anchoring in obscenity. 18 U.S.C. § 1466A criminalizes obscene visual representations of the sexual abuse of children — including drawn, cartoon, and AI-generated imagery — regardless of whether any real child was involved. The constitutional basis is obscenity under Miller v. California, 413 U.S. 15 (1973), which is categorically outside First Amendment protection. Congress enacted § 1466A specifically to close the gap that Ashcroft v. Free Speech Coalition identified in the prior virtual-pornography ban. United States v. Williams, 553 U.S. 285 (2008), upheld the related pandering provisions of § 2252A(a)(3)(B) and is regularly cited as supportive precedent for the broader Adam Walsh framework.
  4. The surviving as-applied Ashcroft question for fully synthetic imagery. Ashcroft v. Free Speech Coalition did not foreclose all virtual-pornography prosecution — it specifically left open the question whether a narrower statute focused on imagery indistinguishable from real children could survive First Amendment review. The current generation of generative-AI imagery is producing fully synthetic images that depict no real person but are perceptually indistinguishable from photographs. Whether § 2256(8)(B) is constitutional as applied to such imagery remains a live as-applied question that is now being litigated in the 2024-2026 prosecution wave and will likely return to the Supreme Court in the next several years.
  5. Texas HB 2700 (2023) closed the state-law gap. Before HB 2700, Texas Penal Code § 43.26 reached only imagery depicting actual children — synthetic imagery fell outside the statute and was not separately criminalized. HB 2700 added subsection (a-1) to reach "visual material that contains a visual depiction of a child younger than 18 years of age engaging in sexual conduct" without requiring that the depicted child be real. The synthetic-imagery prong tracks the same penalty schedule under § 43.26(d): 3rd-degree felony default, 2nd-degree at 6+ items, 1st-degree at 50+ items or victim appearing under 14. Lifetime ch. 62 registration applies to the (a-1) prong as fully as to the (a)(1) prong.
  6. Daubert challenges to forensic AI-detection are the newest defense frontier. AI image-classification tools used in federal CSAM investigations are themselves machine-learning systems with poorly characterized error rates, undisclosed training datasets, and validation processes that have not been subjected to peer review. Federal Rule of Evidence 702 and Daubert require that expert testimony be reliable. A well-crafted Daubert challenge can exclude or limit the government's forensic expert and shift the case from a one-sided forensic presentation to a contested credibility dispute. The defense bar is actively developing this area; appellate authority on the admissibility standards for AI-classification expert testimony is still forming.

An emerging area — state and federal overlap

AI-generated CSAM is the fastest-growing category of state and federal sex-offense prosecution in 2024-2026. Texas Penal Code § 43.26(a-1) (HB 2700) and federal 18 U.S.C. §§ 2252A and 1466A apply concurrently — exposing defendants in the Eastern and Northern Districts of Texas to layered state and federal liability.

Why this area is exploding now
The 2023-2024 wave of consumer-grade generative-AI tools — Stable Diffusion variants, Midjourney, ChatGPT image generation, and open-source diffusion models running on consumer GPUs — has produced a sharp rise in synthetic-imagery prosecutions in federal districts across the country. The Department of Justice and the FBI have publicly identified AI-generated CSAM as a Tier 1 enforcement priority in the National Strategy for Child Exploitation Prevention and Interdiction. State legislatures have followed, with Texas HB 2700 (2023) being the most prominent state-law response to date and similar bills passing in nearly twenty other states between 2023 and 2025.
Concurrent state and federal jurisdiction
A Texas-based defendant who produces, possesses, or distributes AI-generated CSAM is exposed to concurrent prosecution under § 43.26 (state) and §§ 2252A and 1466A (federal). The federal government generally takes the case where interstate-commerce nexus is clear — distribution via the internet, peer-to-peer networks, or social media platforms; production using equipment that traveled in interstate commerce; or possession on devices manufactured outside Texas. The state takes the case where the conduct is purely local or where federal prosecutors decline. Practitioners cannot assume the case will stay in one forum — the case may be referred up to federal court or down to state court at any stage before indictment, and the defense must build for both contingencies from day one.
Federal mandatory minimums change the math
Federal exposure is the dominant strategic factor. Under § 2252A(b)(1), receipt or distribution carries a 5-year mandatory minimum and a 20-year statutory maximum; a prior conviction under chapter 110 (or analogous state law) elevates the mandatory minimum to 15 years and the maximum to 40 years. Production under § 2251 carries a 15-year mandatory minimum and a 30-year maximum (50-year max with one prior). These mandatory minimums apply equally to AI-generated CSAM prosecuted under § 2252A through the § 2256(8)(B) "virtually indistinguishable" pathway. Texas § 43.26 has no mandatory minimum and tops out at 1st-degree felony exposure (5-99 or life), but the federal mandatory minimums regularly make federal court the dispositive forum.
Lifetime sex-offender registration
Conviction under § 43.26 triggers lifetime registration under Texas Code Crim. Proc. ch. 62, including the synthetic-imagery prong under (a-1). Federal conviction under §§ 2252A or 1466A triggers SORNA Tier I/II/III registration under 34 U.S.C. § 20911, with most child-pornography offenses qualifying as Tier I (15-year minimum) or Tier II (25-year minimum). The registration consequence is often more durable than the custodial sentence — a defendant who serves a federal sentence of seven years still faces lifetime or 25-year registration obligations on supervised release and beyond. Registration mitigation is a meaningful component of any AI-CSAM defense, even where guilt itself is not seriously contested.

AI-generated CSAM occupies an unusual position in the current criminal-law landscape: the conduct is being prosecuted aggressively, but the constitutional doctrine governing the prosecutions was written for a world that did not have access to consumer-grade generative-AI imagery. The Supreme Court's 2002 decision in Ashcroft v. Free Speech Coalition struck down the previous virtual-child-pornography ban as overbroad under the First Amendment, holding that the government could not criminalize imagery that depicted no real child and was not obscene under the Miller test. Congress responded the next year with the PROTECT Act, which narrowed the federal definition of "child pornography" to imagery "indistinguishable from" an actual minor under § 2256(8)(B) and added a separate obscenity-based standalone offense under § 1466A. The current generation of federal prosecutions operates through both pathways — sometimes simultaneously charging the same defendant under both statutes — and the defense must understand the structural difference between them to develop a coherent strategy.

The Texas legislature followed Congress's lead in 2023 with HB 2700, which added subsection (a-1) to Penal Code § 43.26 to reach synthetic imagery alongside the long-standing actual-child prong under (a)(1). The result is a tightly integrated state-federal enforcement environment: a defendant arrested in Collin County or Dallas County for possession of AI-generated CSAM faces a state-court 3rd-degree-to-1st-degree felony prosecution under § 43.26 in parallel with potential federal prosecution under § 2252A and/or § 1466A in the Northern or Eastern District of Texas. Practitioners who specialize in either state-court sex-offense defense or federal practice individually are insufficient — these cases require simultaneous fluency in both systems, the federal Sentencing Guidelines framework (USSG § 2G2.2 in particular), and the evolving First Amendment doctrine specific to synthetic imagery.

Federal framework — § 2252A and the "virtually indistinguishable" test

18 U.S.C. § 2252A reaches AI-generated CSAM through § 2256(8)(B), which defines "child pornography" to include any computer-generated image "that is, or is indistinguishable from, that of a minor engaging in sexually explicit conduct." Receipt or distribution carries a 5-year mandatory minimum; production under § 2251 carries 15 years.

The primary federal vehicle for AI-CSAM prosecution is 18 U.S.C. § 2252A, the post-Ashcroft response statute that criminalizes a wide range of conduct relating to "child pornography" as defined under § 2256(8). The offense conduct provisions cover transportation in interstate or foreign commerce (§ 2252A(a)(1)), receipt or distribution (§ 2252A(a)(2)), advertising (§ 2252A(a)(3)), and possession (§ 2252A(a)(5)). The § 2252A(b) penalty structure escalates by conduct type: possession carries a 0-20-year range with no mandatory minimum; receipt or distribution carries a 5-20-year range; production conduct under § 2251 carries 15-30 years. A prior conviction under chapter 110 or analogous state law elevates all of these — receipt or distribution becomes 15-40 years on a single prior, and production exposure can reach 50 years on one prior or life on two priors.

The textual gateway for AI-generated imagery to qualify as "child pornography" for § 2252A purposes is § 2256(8)(B): a "digital image, computer image, or computer-generated image that is, or is indistinguishable from, that of a minor engaging in sexually explicit conduct." The "or is indistinguishable from" clause is the critical language — it does not require that the image actually depict a real child, only that it be indistinguishable from one. The statutory definition of "indistinguishable" in § 2256(11) clarifies that the image must be one that an "ordinary person viewing the depiction would conclude" depicts an actual minor — a perceptual test that places the question squarely in front of the jury and that has become the central evidentiary battle in modern AI-CSAM prosecutions.

The "virtually indistinguishable" framework creates two distinct litigation fronts in any § 2252A AI-CSAM case. First, the factual front: can the government prove that the specific imagery at issue is, in fact, indistinguishable from an actual minor in the perception of an ordinary person? This is the domain of forensic AI-detection expert testimony, jury demonstratives, and increasingly the use of generative-AI metadata embedded in the image files themselves. Second, the constitutional front: is § 2256(8)(B) constitutional as applied to fully synthetic imagery with no real-child basis? Ashcroft did not foreclose all virtual-pornography prosecution — it specifically left open the question whether a narrower statute focused on imagery indistinguishable from real children could survive First Amendment review. That question has been litigated in lower courts since 2003 but has not returned to the Supreme Court, and the rise of high-quality generative-AI imagery is producing a new generation of as-applied challenges in 2024-2026.

United States v. Mecham, 950 F.3d 257 (5th Cir. 2020), is the Fifth Circuit's most direct decision on the sufficiency of the evidence for an AI-related child-pornography conviction — though Mecham predates the current generative-AI wave. The court there upheld the conviction on the basis that the jury could reasonably conclude that the imagery in question depicted real minors notwithstanding the defendant's claim of digital alteration. The case illustrates the deference appellate courts give to jury findings on the "actual minor" versus "computer-generated" question, and the difficulty of overturning a conviction on insufficiency grounds where the government has produced any evidence — testimonial, forensic, or perceptual — supporting the conclusion that the depicted minor is real. United States v. Spencer, 113 F.4th 1129 (8th Cir. 2024), is the most prominent recent decision addressing AI-CSAM prosecution and is now widely cited as the leading appellate authority on the application of § 2252A and the relevant Sentencing Guidelines provisions to generative-AI imagery.

Federal framework — § 1466A obscenity-based standalone CSAM

Section 1466A is the obscenity-based companion to § 2252A — it reaches drawn, cartoon, and AI-generated imagery regardless of whether any real child was involved. The constitutional basis is obscenity under Miller v. California, not the harm-to-real-children rationale of § 2252A. Penalties match § 2252A.

18 U.S.C. § 1466A, enacted as part of the PROTECT Act of 2003, is the second federal pathway for AI-CSAM prosecution and the one with the deepest constitutional pedigree. Section 1466A criminalizes the production, distribution, receipt, or possession of either (1) "a visual depiction of any kind, including a drawing, cartoon, sculpture, or painting, that depicts a minor engaging in sexually explicit conduct" that is obscene, or (2) such a visual depiction that "lacks serious literary, artistic, political, or scientific value." The "minor" in § 1466A need not be a real person — the statute reaches purely fictional or synthetic depictions, including AI-generated imagery, where the constitutional category of obscenity applies. The penalty structure mirrors § 2252A: a 5-year mandatory minimum for distribution, escalating with priors, and similar production exposure under § 2251.

The structural advantage of § 1466A from the prosecutor's perspective is that it bypasses the constitutional question lurking in § 2252A as applied to fully synthetic imagery. Ashcroft v. Free Speech Coalition turned on the First Amendment principle that the government cannot criminalize imagery merely because it depicts a minor engaged in sexually explicit conduct, where no actual minor was harmed in the production and the imagery is not obscene under Miller. Section 1466A sidesteps that limitation by anchoring liability in obscenity itself — and obscenity has been categorically outside First Amendment protection since Roth v. United States, 354 U.S. 476 (1957), and Miller v. California, 413 U.S. 15 (1973). If the imagery at issue meets the three-part Miller test (appeals to prurient interest, depicts sexual conduct in a patently offensive way, lacks serious literary/artistic/political/scientific value as determined by contemporary community standards), prosecution under § 1466A is constitutionally permissible regardless of whether any real child was depicted.

United States v. Williams, 553 U.S. 285 (2008), upheld the related pandering provisions in § 2252A(a)(3)(B) — the provision criminalizing offers to provide or solicit child pornography — against First Amendment challenge. Williams did not directly address § 1466A, but it is regularly cited as supportive precedent for the broader proposition that Congress retains substantial authority to criminalize conduct related to the production and distribution of child sexual abuse material, including conduct that does not necessarily involve actual children. Lower-court applications of § 1466A have generally upheld convictions involving drawn, cartoon, and synthetic imagery where the obscenity element was supported by the trial record.

The charge-bargaining significance of § 1466A versus § 2252A is substantial. A defendant facing a § 2252A indictment for AI-CSAM may be able to negotiate a plea down to § 1466A, or vice versa, depending on the relative strength of the government's proof on the "virtually indistinguishable" element versus the Miller obscenity elements. The penalty structures are similar, but the collateral consequences — including the precise SORNA tier classification and the language that appears on the public registry — can vary in ways that matter to post-incarceration life. Experienced federal defense counsel evaluate both pathways at the earliest possible stage and position the case for the most favorable charge framework.

The constitutional trajectory — Ashcroft → Adam Walsh → Williams

The current statutory framework is the product of a three-step constitutional dialogue: Ashcroft v. Free Speech Coalition (2002) struck down the prior virtual-pornography ban; the PROTECT Act / Adam Walsh framework (2003) responded with the "virtually indistinguishable" test and § 1466A; United States v. Williams (2008) upheld the resulting pandering provisions.

Understanding the current federal framework requires understanding the constitutional dialogue that produced it. Before 2003, the federal definition of "child pornography" included any visual depiction that "is, or appears to be, of a minor engaging in sexually explicit conduct" — former 18 U.S.C. § 2256(8)(D). The Free Speech Coalition challenged that provision on First Amendment grounds, arguing that the "appears to be" language reached imagery that depicted no real child and was not obscene under Miller. In Ashcroft v. Free Speech Coalition, 535 U.S. 234 (2002), the Supreme Court agreed and struck down § 2256(8)(D) as substantially overbroad. The Court emphasized that the government's established interest in protecting actual children from sexual exploitation could not extend to a categorical ban on virtual imagery that depicted no real victim and was not obscene.

The Court's opinion left open two pathways for narrower legislation. First, the Court did not foreclose prosecution of virtual imagery that was indistinguishable from real children — the opinion repeatedly emphasized that the "appears to be" language swept too broadly precisely because it reached imagery that was not indistinguishable from actual minors. Second, the Court explicitly noted that obscene imagery falls outside First Amendment protection regardless of whether it depicts a real child, leaving the government free to enact obscenity-based statutes reaching virtual imagery. Congress responded to both invitations the following year through the PROTECT Act of 2003, which (1) replaced the "appears to be" language with the "or is indistinguishable from" framework now in § 2256(8)(B) and (2) enacted § 1466A as a standalone obscenity-based offense.

United States v. Williams, 553 U.S. 285 (2008), provided the Supreme Court's next significant intervention. The Williams Court upheld § 2252A(a)(3)(B), the post-Ashcroft pandering provision that criminalizes offers to provide or solicit material represented or believed to be child pornography. The Court applied a "speech integral to criminal conduct" rationale, holding that offers to engage in illegal transactions categorically fall outside First Amendment protection. Williams did not address § 2256(8)(B) directly, and the Court has not since revisited the constitutionality of the "virtually indistinguishable" framework as applied to fully synthetic AI-generated imagery — the question that the 2024-2026 prosecution wave is now pressing into the lower courts.

The surviving constitutional question — and the most important as-applied First Amendment issue in any AI-CSAM defense — is whether § 2256(8)(B) is constitutional as applied to fully synthetic imagery with no real-child basis. Ashcroft's logic suggests that purely synthetic imagery cannot be criminalized merely because it depicts a fictional minor engaged in sexually explicit conduct, where no actual child was harmed and the imagery is not obscene. Congress's response was that imagery "indistinguishable from" a real child is, perceptually if not actually, a depiction of an actual minor and is therefore within the scope of the government's established child-protection interest. Whether the Supreme Court would accept that perceptual-indistinguishability rationale as a sufficient hook for criminal prosecution of AI imagery — at a moment when generative AI is producing images that are, in fact, indistinguishable from photographs of real people who do not exist — is the live constitutional question that the next wave of federal appellate review will likely confront.

Texas framework — § 43.26 and the HB 2700 expansion

Texas Penal Code § 43.26(a)(1) has long criminalized possession or promotion of child pornography; HB 2700 (2023) added subsection (a-1) to reach synthetic and AI-generated imagery on the same penalty schedule. Default 3rd-degree felony elevates to 2nd-degree at 6+ items, 1st-degree at 50+ items or victim under 14.

Texas Penal Code § 43.26 is the primary state-court vehicle for child pornography possession and promotion offenses, and HB 2700 — passed by the 88th Texas Legislature in 2023 — extended its reach to AI-generated and other synthetic imagery. Subsection (a)(1) continues to govern conduct involving visual material that depicts an actual child younger than 18, and subsection (a-1) — the HB 2700 addition — reaches "visual material that contains a visual depiction of a child younger than 18 years of age engaging in sexual conduct" without requiring that the depicted child be real. The legislative history of HB 2700 made the synthetic-imagery purpose explicit: the bill author repeatedly identified AI-generated CSAM as the targeted conduct, and the statutory text was drafted with both AI imagery and deepfake face-swap conduct in mind.

The penalty grading under § 43.26(d) operates by volume of material rather than by whether the imagery is actual or synthetic. The default offense is a 3rd-degree felony (2-10 years in TDCJ and a $10,000 fine). Possession of 6 or more items elevates the offense to a 2nd-degree felony (2-20 years and $10,000 fine). Possession of 50 or more items, or possession of imagery depicting a child under 14, elevates the offense to a 1st-degree felony (5-99 years or life and a $10,000 fine). The "item" count is determined by individual visual file in most charging practice, though aggregating practices vary by district attorney's office. Promotion conduct (distribution, sale, exhibition) under § 43.26(e) is enhanced relative to mere possession.

Texas registration consequences are severe and largely overlap with federal SORNA requirements. Conviction under § 43.26 — including the synthetic-imagery prong under (a-1) — is a "reportable conviction" under Code Crim. Proc. art. 62.001 and triggers lifetime registration under art. 62.101. The registration obligation persists throughout life unless statutory termination criteria are satisfied; the typical defendant convicted of § 43.26 will remain on the registry indefinitely. Public registry entries include the conviction offense, registration tier, photograph, address, employment, vehicle, and other identifying information. The collateral consequences of registration — employment restrictions, housing restrictions, travel notifications, school-proximity prohibitions — are often more disruptive to post-incarceration life than the custodial sentence itself.

Affirmative defenses under § 43.26 are narrow. Subsection (f) creates an affirmative defense for possession of material that the defendant promptly destroyed or reported to law enforcement and that the defendant did not solicit or knowingly distribute. The ten-item storage cap defense applicable to certain other Texas obscenity offenses does not apply to § 43.26. The principal defense moves in a state-court § 43.26 case are therefore the same moves available in a federal § 2252A or § 1466A case: suppression of digital evidence under the Fourth Amendment, mens-rea contests on the knowledge element, First Amendment as-applied challenges to (a-1) for fully synthetic imagery, and forensic challenges to the State's identification of the imagery as depicting a minor.

Defense strategies for AI-CSAM cases

AI-CSAM defense work centers on First Amendment as-applied challenges, Fourth Amendment suppression of digital evidence, mens-rea contests on the knowledge-of-minor element, Daubert challenges to forensic AI-detection expert testimony, and charge-bargaining between § 2252A and § 1466A.

The defense playbook in an AI-CSAM case is structurally different from a traditional CSAM defense because the threshold legal questions are themselves contested. In a traditional § 43.26 or § 2252A case involving imagery depicting a real child, the principal defense moves are mens-rea (knowing possession) and Fourth Amendment suppression. In an AI-CSAM case, those moves remain available, but they are joined by First Amendment as-applied challenges (does § 2256(8)(B) reach this imagery constitutionally?), forensic identification contests (can the State actually prove this image is indistinguishable from a real minor?), and statutory interpretation arguments (does the conduct fall within (a-1) or only within (a)(1)?). The result is a denser pretrial motion practice than in most state-court sex-offense cases.

First Amendment as-applied challenges to § 2256(8)(B) and to § 43.26(a-1) are the most ambitious — and potentially the most consequential — defense move. The argument runs through Ashcroft v. Free Speech Coalition: where the imagery at issue was generated entirely by a generative-AI system with no real-child basis, where no real child was harmed in the production, and where the imagery is not obscene under the Miller test, the First Amendment forbids criminal prosecution. The government's response is the "virtually indistinguishable" framework — that perceptual indistinguishability provides the constitutional hook. The litigation is fact-bound and turns heavily on the specific imagery at issue, the production methodology, and the available expert testimony on identifiability. United States v. Spencer, 113 F.4th 1129 (8th Cir. 2024), is now the leading appellate authority on these issues.

Fourth Amendment suppression remains the highest-value pretrial move in most AI-CSAM cases. The federal investigative pattern for these cases routinely involves border searches of devices, tip-driven warrants for cloud accounts, peer-to-peer monitoring software, and increasingly the use of AI-based forensic tools to identify suspect imagery on seized devices. Each investigative step is a suppression opportunity. Riley v. California, 573 U.S. 373 (2014), governs warrant requirements for cell-phone searches. Carpenter v. United States, 585 U.S. 296 (2018), restricts warrantless cell-site location information acquisition. Pope v. State, 685 S.W.3d 39 (Tex. Crim. App. 2024), is the recent Texas decision on suppression-of-electronic-evidence standards. Particularity, staleness, scope-of-search, and good-faith reliance issues are routine in these cases.

Mens-rea contests on the knowing-possession element have become more important in AI-CSAM cases than in traditional CSAM. The government must prove the defendant knew the imagery depicted a minor (or imagery indistinguishable from a minor) — and in AI-CSAM cases, defendants frequently raise the question whether they believed the depicted person was an adult or whether they knew the image was AI-generated at all. The credibility of these defenses varies by case, but they are no longer unusual. The government typically responds with file-name analysis, search-history evidence, and circumstantial proof of knowledge. The defense responds with cross-examination of forensic examiners, character testimony, and where appropriate, expert testimony on the limits of AI image classification.

Daubert challenges to forensic AI-detection expert testimony are the newest and most rapidly evolving area of AI-CSAM defense. Federal Rule of Evidence 702 and Daubert require that expert testimony be the product of reliable principles and methods reliably applied to the facts of the case. AI image-classification tools are themselves machine-learning systems, often trained on undisclosed datasets, with error rates that are poorly characterized and validation processes that have not been subjected to peer review. A well-crafted Daubert challenge can exclude or limit the government's forensic expert and shift the case from a one-sided forensic presentation to a contested credibility dispute. The defense bar is actively developing this area; the federal courts have not yet uniformly addressed the admissibility standards for AI-classification expert testimony.

Charge-bargaining between § 2252A and § 1466A is a structurally important strategic lever. Where the government's proof on the "virtually indistinguishable" element is weak, defense counsel can sometimes negotiate a plea to § 1466A (obscenity-based) in lieu of § 2252A — particularly where the imagery is plainly synthetic and the obscenity elements are easier to establish. Where the imagery is closer to real-child indistinguishability and the obscenity elements are weaker, the negotiation may run in the opposite direction. The penalty structures are similar, but the SORNA tier classification and the precise language on the registry can differ, and the federal Sentencing Guidelines applications under § 2G2.2 versus § 2G3.1 differ in important respects.

Penalty exposure — lifetime registration and federal mandatory minimums

Federal § 2252A imposes a 5-year mandatory minimum for receipt or distribution, 15 years on a prior conviction, and 15 years for § 2251 production. Texas § 43.26 carries no mandatory minimum but reaches 1st-degree felony exposure (5-99 or life). Both trigger lifetime sex-offender registration.

The penalty exposure in AI-CSAM cases is dominated by the federal mandatory minimums. Section 2252A(b)(1) imposes a 5-year mandatory minimum for receipt or distribution and a 15-year mandatory minimum on a prior chapter-110 conviction. Production conduct under § 2251 carries a 15-year mandatory minimum, escalating to 25 years on one prior and 35 years on two priors. The statutory maximums are 20 years (possession or receipt without prior), 40 years (receipt/distribution with one prior), and 30-50 years or life (production with priors). The Sentencing Guidelines under USSG § 2G2.2 (possession/receipt/distribution) and § 2G2.1 (production) typically produce advisory ranges substantially above the statutory minimum — driven by use of a computer enhancement (+2), number-of-images enhancements (+2 to +5), prepubescent-minor enhancement (+2), and sadistic-or-violent enhancement (+4).

Federal sentencing in CSAM cases — including AI-CSAM — is heavily influenced by the post-Booker variance pattern in non-production CSAM cases. Substantial numbers of federal district judges have been varying downward from the § 2G2.2 advisory range for over a decade on the theory that the enhancements compound in ways that produce excessive sentences in run-of-the-mill possession cases. The variance pattern is less pronounced in distribution and production cases and even less so in cases involving sadistic, sadomasochistic, or violent imagery. The strategic implication for defense planning: the advisory range is the starting point, not the endpoint, and a well-developed § 3553(a) mitigation presentation focused on individual characteristics, treatment, and the relative culpability of the conduct can produce sentences materially below the advisory range in possession and receipt cases.

Texas § 43.26 has no mandatory minimum but produces serious exposure under § 12.32-12.34. The default 3rd-degree felony carries 2-10 years in TDCJ; the 6+ items elevation produces a 2nd-degree felony with 2-20 years; the 50+ items or under-14 elevation produces a 1st-degree felony with 5-99 years or life. Community supervision is available on the 3rd-degree and (in jury-recommended scenarios with the eligibility criteria met) the 2nd-degree variants; 1st-degree exposure is structurally less amenable to probation but remains theoretically available in narrow circumstances. Parole eligibility runs under Gov't Code § 508.145, with no 3g aggravated-offense designation specific to § 43.26 — though the inevitable deadly-weapon or "sexually violent offense" overlay can shift the calculation.

Texas Code Crim. Proc. ch. 62 lifetime registration is the universal collateral consequence of any § 43.26 conviction. Registration runs under art. 62.101(a) for the duration of the defendant's life unless statutory termination criteria are met. Federal SORNA registration under 34 U.S.C. § 20911 et seq. operates concurrently with state registration and is required as a condition of supervised release in federal cases. Most § 2252A and § 1466A convictions classify as SORNA Tier I (15-year minimum), Tier II (25-year minimum), or Tier III (lifetime), with the tier dependent on the specific offense subsection, the age of the depicted minor, and prior convictions. Texas registry residency restrictions under Loc. Gov't Code § 341.906 and analogous municipal ordinances apply to all registrants and significantly constrain post-incarceration housing options across DFW.

Local DFW practice — HSI, ICAC, FBI Cyber, NDOTX prosecutorial focus

AI-CSAM cases in DFW are typically investigated by HSI Homeland Security Investigations, the North Texas ICAC Task Force, and the FBI Dallas Cyber Squad. Federal prosecution proceeds in the Northern District of Texas (Dallas Division) and the Eastern District of Texas (Sherman/Plano divisions).

The federal investigative apparatus for AI-CSAM cases in DFW is concentrated in three primary agencies. Homeland Security Investigations (HSI) operates the Dallas Field Office Cyber Crimes Center and runs the largest single docket of federal CSAM investigations in the region. The North Texas Internet Crimes Against Children (ICAC) Task Force is housed at the Dallas Police Department and coordinates between federal, state, and local investigators across Collin, Dallas, Denton, and Tarrant Counties. The FBI Dallas Cyber Squad handles the federal side of investigations originating from National Center for Missing & Exploited Children (NCMEC) CyberTipline referrals, internet-service-provider reports, and undercover-operation tips. The three agencies coordinate closely, and a case that starts with one routinely involves the others by indictment stage.

Federal prosecution in DFW is split between the Northern District of Texas and the Eastern District of Texas. The Northern District (Dallas Division) handles cases arising in Dallas and surrounding counties west and south; the Eastern District (Sherman Division, Plano courthouse) handles cases arising in Collin, Denton, Grayson, and Fannin Counties. The two districts have separate U.S. Attorney's Offices, separate Federal Public Defender offices, and somewhat different charging and pleading practices. The Eastern District of Texas (NDOTX is the older abbreviation; the current standard abbreviation is EDTX, with NDTX for Northern) has historically taken an aggressive prosecutorial posture in CSAM cases, with high rates of indictment, high plea rates, and a generally guideline-anchored sentencing pattern. Defense counsel without significant federal practice experience in these specific districts are at a substantial disadvantage.

State-court prosecutions of § 43.26 in DFW are handled by the District Attorney's Offices in Collin County (McKinney), Dallas County (Dallas), Denton County (Denton), and Tarrant County (Fort Worth). Each office has a specialized prosecutorial unit for sex-offense and crimes-against-children cases. The state-court venue selection often follows the conduct location, but in cases that originate from a CyberTipline referral or a forensic-imagery analysis, the venue can be flexible. The interplay between state and federal prosecutors in these cases is complex — a CyberTipline referral that comes through ICAC may be charged state-court if the federal prosecutor declines, may be charged federally if the state prosecutor defers, or may be charged in both forums in cases involving distinct conduct. Defense counsel must be alert to the dual-sovereignty doctrine and to the practical pattern by which DFW prosecutors coordinate (or fail to coordinate) charging decisions.

L and L Law Group's practice in this area covers state-court defense in all four DFW counties and federal defense in both the Northern and Eastern Districts of Texas. The firm's federal practice includes admissions to TXND, TXED, and the Fifth Circuit. Co-Founding Partners Reggie London (Tex. Bar 24043514) and Njeri London (Tex. Bar 24043266) lead the firm's sex-offense and digital-evidence work. The firm represents defendants from initial law-enforcement contact through grand jury, indictment, pretrial motion practice, trial, sentencing, and post-conviction proceedings. The earlier the firm is engaged in an investigation, the more leverage is available — including pre-indictment negotiations with the U.S. Attorney's Office, target letter response, and Fifth Amendment counseling on subpoena response. Contact (972) 370-5060 for confidential consultation.

Defense Strategy

What we evaluate first

Five defense levers do most of the work in Texas evading cases. We evaluate every one before charting a path — suppression first, then knowledge, intent, necessity, and charge-reduction posture together set the strategy.

  1. First Amendment as-applied challenge for fully synthetic imagery
    Where the imagery at issue was generated entirely by a generative-AI system with no real-child basis, where no real child was harmed in production, and where the imagery is not obscene under the Miller test, the First Amendment forbids prosecution under § 2256(8)(B) — that is the surviving Ashcroft question. The argument runs through Ashcroft v. Free Speech Coalition, 535 U.S. 234 (2002), and the lower-court applications since 2003. The litigation is fact-bound and turns on the specific imagery, production methodology, and available expert testimony on whether the imagery is in fact "indistinguishable from" an actual minor in the perception of an ordinary person. The challenge has the highest dismissal potential where production methodology is documentable and the imagery is clearly synthetic.
  2. § 1466A vs. § 2252A charge-bargaining
    A defendant facing a § 2252A indictment for AI-CSAM may be able to negotiate a plea down to § 1466A (obscenity-based) — or vice versa — depending on the relative strength of the government's proof on the "virtually indistinguishable" element versus the Miller obscenity elements. The penalty structures are similar, but the SORNA tier classification, the precise language on the public registry, and the Sentencing Guidelines applications under § 2G2.2 versus § 2G3.1 differ. Experienced federal defense counsel evaluate both pathways at the earliest possible stage and position the case for the more favorable charge framework. Where the imagery is plainly synthetic, § 1466A may be the more favorable charge; where the imagery is closer to real-child indistinguishability, § 2252A may be.
  3. Fourth Amendment suppression of digital evidence
    Federal AI-CSAM investigations routinely involve border searches of devices, tip-driven warrants for cloud accounts, peer-to-peer monitoring software, and AI-based forensic tools applied to seized devices. Each is a suppression opportunity. Riley v. California, 573 U.S. 373 (2014), requires a warrant for cell-phone searches; Carpenter v. United States, 585 U.S. 296 (2018), restricts warrantless cell-site location information acquisition; Pope v. State, 685 S.W.3d 39 (Tex. Crim. App. 2024), is the recent Texas decision on suppression-of-electronic-evidence standards. Particularity, staleness, scope-of-search, and Leon good-faith reliance issues are routine. Suppression of the seized imagery is often case-dispositive.
  4. Mens-rea contests on knowledge of minor
    The government must prove the defendant knew the imagery depicted a minor — or, in AI-CSAM cases, knew the imagery was indistinguishable from a minor. Defendants frequently raise the question whether they believed the depicted person was an adult, whether they knew the image was AI-generated, or whether they knew an apparently adult person depicted was actually a minor. The government typically responds with file-name analysis, search-history evidence, and circumstantial proof. The defense responds with cross-examination of forensic examiners, character testimony, and where appropriate, expert testimony on the limits of AI image classification. The mens-rea defense is more credible in AI cases than in traditional CSAM because the production tools themselves obscure subject identification.
  5. Daubert challenge to forensic AI-detection expert testimony
    Federal Rule of Evidence 702 and Daubert v. Merrell Dow Pharmaceuticals, 509 U.S. 579 (1993), require that expert testimony be the product of reliable principles and methods reliably applied. AI image-classification tools used in federal investigations are themselves machine-learning systems with poorly characterized error rates, undisclosed training datasets, and validation processes that have not been peer-reviewed. A well-crafted Daubert challenge can exclude or limit the government's forensic expert. The defense bar is actively developing this area and appellate authority on admissibility standards for AI-classification expert testimony is still forming — meaning early adopters of these challenges may shape the controlling precedent in their respective circuits.
  6. Entrapment defense in undercover sting operations
    A meaningful portion of federal AI-CSAM cases arise from undercover sting operations — typically through peer-to-peer monitoring or social-media platform decoys. Where the government induced the defendant to commit conduct he was not predisposed to commit, the entrapment defense under Jacobson v. United States, 503 U.S. 540 (1992), and Mathews v. United States, 485 U.S. 58 (1988), is available. The defense focuses on government inducement, the defendant's lack of predisposition, and (where the facts support it) the absence of any prior conduct suggesting prurient interest in minors. Texas entrapment under PC § 8.06 is structurally similar and applies in state-court prosecutions. The defense is fact-intensive but is one of the few that can produce outright acquittal in an undercover-sting case.
  7. Affirmative defense — § 43.26 prompt-destruction or reporting
    Texas Penal Code § 43.26(f) creates a limited affirmative defense for possession of material that the defendant promptly destroyed or reported to law enforcement and that the defendant did not solicit or knowingly distribute. The defense is narrow but available — it requires that the destruction or reporting be prompt (typically within hours or a small number of days of receipt) and that the defendant not have engaged in any conduct inconsistent with non-solicitation and non-distribution. The defense has no precise federal analog; in federal court the parallel question is sometimes raised as a sentencing mitigation under § 3553(a) rather than as an affirmative defense at trial.
Defense Timeline

How we build the case

Texas evading defense follows a predictable four-phase arc — stabilize and discover (0-15 days), build the suppression record (15-90 days), motion practice and posture (3-6 months), then trial readiness or resolution (6 months+).

  1. Day 0-30 (Investigation)
    Pre-arrest investigation and target-letter posture
    Federal AI-CSAM cases routinely begin with a CyberTipline referral, an ISP report, or a peer-to-peer monitoring detection followed by weeks or months of investigative buildup before the defendant becomes aware. Where the defendant learns of the investigation via target letter, grand jury subpoena, or witness contact, immediate engagement of federal-experienced counsel is essential. Pre-arrest steps: invoke Fifth Amendment on any contact; preserve and isolate devices (no destruction — that creates separate obstruction exposure); consider pre-indictment negotiations with the U.S. Attorney's Office; document AI-tool production methodology where the defense theory will rest on synthetic imagery; engage a forensic-imagery expert early.
  2. Day 30-90 (Indictment)
    Grand jury, indictment, and discovery
    Federal indictment typically follows several months of investigative buildup. Initial appearance, detention hearing, and arraignment occur in rapid succession. Discovery under Rule 16 produces the seized devices, forensic reports, AI-classification expert reports, and any law-enforcement interviews. Forensic-imagery and AI-detection expert retention by defense at the earliest stage. State-court parallel proceedings (if any) coordinate with federal counsel — sometimes defending in both forums simultaneously, sometimes positioning for federal dismissal in favor of state-court resolution. Bond posture is critical — pretrial detention is the rule, not the exception, in federal CSAM cases under 18 U.S.C. § 3142(e)(3)(E).
  3. Month 3-12 (Pretrial)
    Motion practice — suppression, First Amendment, Daubert
    Pretrial motion practice is the highest-leverage phase of an AI-CSAM defense. Suppression motions challenge the underlying device searches, cloud account warrants, peer-to-peer monitoring, and forensic-tool methodology. First Amendment as-applied motions under Ashcroft v. Free Speech Coalition challenge § 2256(8)(B) for fully synthetic imagery. Daubert motions challenge the AI-classification forensic expert testimony. Rule 12 motions to dismiss for charging defects, particularity issues, or constitutional vagueness. The volume and density of motion practice in these cases is substantially higher than in traditional CSAM cases, and the outcome of pretrial motion practice often determines whether the case resolves at plea or proceeds to trial.
  4. Month 12+ (Resolution)
    Trial or plea, sentencing, and post-conviction
    Trial settings typically 12-24 months from indictment. Trial proceeds with bifurcated guilt-then-sentencing structure in federal court. Sentencing in federal CSAM cases is dominated by the § 2G2.2 (possession/receipt/distribution) or § 2G2.1 (production) Sentencing Guidelines calculation, with § 3553(a) variance arguments for non-production cases producing meaningful below-guideline sentences in a substantial portion of cases. Lifetime sex-offender registration under SORNA and Texas ch. 62 attaches automatically on conviction. Post-conviction representation includes § 2255 collateral attack, registration-related litigation, and (in some cases) executive clemency advocacy.

Charged with evading arrest in Collin, Denton, Dallas, or Tarrant County?

L and L Law Group defends evading-arrest cases at every level — misdemeanor through second-degree felony. Free initial consultation.

Call (972) 370-5060

Frequently asked questions

Twelve questions we answer most often about Texas evading-arrest cases — penalties, defenses, expunction, court timeline, license impact, and federal-case interaction.

Is AI-generated child sexual abuse material illegal in Texas?

Yes. Texas Penal Code § 43.26(a-1), added by HB 2700 in 2023, criminalizes the possession or promotion of "visual material that contains a visual depiction of a child younger than 18 years of age engaging in sexual conduct" without requiring that the depicted child be real. The synthetic-imagery prong runs parallel to the actual-child prong under § 43.26(a)(1) and is graded on the same penalty schedule under § 43.26(d): 3rd-degree felony default (2-10 years and a $10,000 fine), 2nd-degree felony at 6 or more items (2-20 years), and 1st-degree felony at 50 or more items or where the depicted child appears under 14 (5-99 years or life). Lifetime sex-offender registration under Code Crim. Proc. ch. 62 applies on conviction.

Is AI-generated CSAM illegal under federal law?

Yes — through two distinct statutory pathways. First, 18 U.S.C. § 2252A criminalizes a wide range of conduct relating to "child pornography" as defined under § 2256(8)(B), which expressly includes any "computer-generated image that is, or is indistinguishable from, that of a minor engaging in sexually explicit conduct." Second, 18 U.S.C. § 1466A criminalizes obscene visual representations of the sexual abuse of children — including drawn, cartoon, and AI-generated imagery — regardless of whether any real child was depicted. Federal prosecutors routinely charge under either or both statutes depending on the imagery and the available proof. Receipt or distribution under § 2252A carries a 5-year mandatory minimum; production under § 2251 carries a 15-year mandatory minimum.

What is the "virtually indistinguishable" test in 18 U.S.C. § 2256(8)(B)?

18 U.S.C. § 2256(8)(B) defines "child pornography" to include any "digital image, computer image, or computer-generated image that is, or is indistinguishable from, that of a minor engaging in sexually explicit conduct." The statutory definition of "indistinguishable" in § 2256(11) clarifies that the image must be one that an "ordinary person viewing the depiction would conclude" depicts an actual minor — a perceptual test placed in front of the jury. Congress enacted this language in the PROTECT Act of 2003 after the Supreme Court struck down the prior virtual-pornography ban in Ashcroft v. Free Speech Coalition, 535 U.S. 234 (2002). The "virtually indistinguishable" framework is the textual gateway by which AI-generated imagery becomes "child pornography" for § 2252A purposes.

What did Ashcroft v. Free Speech Coalition hold, and does it still apply to AI-generated imagery?

Ashcroft v. Free Speech Coalition, 535 U.S. 234 (2002), struck down former 18 U.S.C. § 2256(8)(D), which had defined "child pornography" to include any image that "appears to be" a minor engaging in sexually explicit conduct. The Court held that the "appears to be" language swept too broadly under the First Amendment because it criminalized imagery that depicted no real child and was not obscene under the Miller test. Congress responded the following year with the PROTECT Act of 2003, which replaced "appears to be" with the narrower "virtually indistinguishable" language now in § 2256(8)(B) and enacted § 1466A as an obscenity-based standalone offense. Whether § 2256(8)(B) is constitutional as applied to fully synthetic AI imagery — imagery with no real-child basis — remains a live as-applied question being litigated in the 2024-2026 prosecution wave.

What is the difference between § 2252A and § 1466A?

Both statutes criminalize CSAM-related conduct, but they rest on different constitutional foundations. Section 2252A operates through § 2256(8)'s definition of "child pornography" — including the (B) "virtually indistinguishable" pathway for AI imagery — and rests on the government's established interest in protecting actual children. Section 1466A operates through obscenity doctrine under Miller v. California, 413 U.S. 15 (1973), and reaches drawn, cartoon, and AI-generated imagery even where no real child was depicted. The penalty structures are similar — both impose a 5-year mandatory minimum for distribution and 15-year mandatory minimum for production — but the SORNA tier classification, the precise registry language, and the Sentencing Guidelines applications (§ 2G2.2 versus § 2G3.1) differ. Charge-bargaining between the two is a meaningful defense strategy in synthetic-imagery cases.

What are the federal mandatory minimums for AI-CSAM offenses?

18 U.S.C. § 2252A(b)(1) imposes a 5-year mandatory minimum for receipt or distribution of child pornography (including AI-generated imagery that qualifies under § 2256(8)(B)) and a 20-year statutory maximum. A prior conviction under chapter 110 or analogous state law elevates the mandatory minimum to 15 years and the maximum to 40 years. Production under 18 U.S.C. § 2251 carries a 15-year mandatory minimum and a 30-year maximum, escalating to 25 years on one prior and 35 years on two priors. Mere possession under § 2252A(b)(2) does not carry a mandatory minimum but exposes the defendant to 0-20 years. Section 1466A mirrors these mandatory minimums in most respects. Texas § 43.26 has no mandatory minimum but reaches 1st-degree felony exposure (5-99 or life) at the upper grading tiers.

What are the Texas penalties for AI-generated CSAM under § 43.26?

Texas Penal Code § 43.26(d) grades the offense by volume of material rather than by whether the imagery is actual or synthetic. The default offense is a 3rd-degree felony (2-10 years in TDCJ and a $10,000 fine under § 12.34). Possession of 6 or more items elevates the offense to a 2nd-degree felony (2-20 years and $10,000 fine under § 12.33). Possession of 50 or more items, or possession of imagery depicting a child who appears under 14, elevates the offense to a 1st-degree felony (5-99 years or life and a $10,000 fine under § 12.32). Promotion conduct (distribution, sale, exhibition) under § 43.26(e) is enhanced relative to mere possession. Lifetime sex-offender registration under Code Crim. Proc. ch. 62 applies on any conviction.

Can I raise a First Amendment defense to an AI-CSAM charge?

Yes — for imagery that is fully synthetic with no real-child basis, an as-applied First Amendment challenge to 18 U.S.C. § 2256(8)(B) is available. The argument runs through Ashcroft v. Free Speech Coalition, 535 U.S. 234 (2002): where no real child was harmed in production, the imagery is not obscene under the Miller test, and the depicted "minor" does not exist, the First Amendment forbids criminal prosecution. The government's response is the "virtually indistinguishable" framework — that perceptual indistinguishability from an actual minor provides the constitutional hook. The challenge is most viable where the production methodology is documentable (the defendant can prove the image was AI-generated and no real person was the source) and where the imagery does not satisfy the Miller obscenity elements. The defense is structurally unavailable against a § 1466A obscenity-based charge if the imagery satisfies Miller.

What is the role of Fourth Amendment suppression in AI-CSAM defense?

Fourth Amendment suppression is the highest-value pretrial move in most federal and state AI-CSAM cases. Federal investigations routinely involve border searches of devices, tip-driven warrants for cloud accounts, peer-to-peer monitoring software, and AI-based forensic tools applied to seized devices — each step is a suppression opportunity. Riley v. California, 573 U.S. 373 (2014), requires a warrant for cell-phone searches incident to arrest. Carpenter v. United States, 585 U.S. 296 (2018), requires a warrant for cell-site location information. Pope v. State, 685 S.W.3d 39 (Tex. Crim. App. 2024), is the recent Texas decision on suppression-of-electronic-evidence standards. Successful suppression of the seized imagery is frequently case-dispositive because there is rarely an alternative source for the underlying evidence.

How does sex-offender registration work for an AI-CSAM conviction?

Conviction under Texas Penal Code § 43.26 — including the synthetic-imagery prong under (a-1) — is a "reportable conviction" under Code Crim. Proc. art. 62.001 and triggers lifetime registration under art. 62.101. Federal SORNA registration under 34 U.S.C. § 20911 et seq. operates concurrently with state registration and is required as a condition of supervised release in federal cases. Most § 2252A and § 1466A convictions classify as SORNA Tier I (15-year minimum), Tier II (25-year minimum), or Tier III (lifetime), with the tier dependent on the specific offense subsection, the age of the depicted minor, and prior convictions. Public registry entries include the conviction offense, registration tier, photograph, address, employment, vehicle, and other identifying information. Residency restrictions under Loc. Gov't Code § 341.906 and analogous municipal ordinances significantly constrain post-incarceration housing options across DFW. Registration is often more durable than the custodial sentence itself.

Which agencies investigate AI-CSAM cases in DFW?

Three primary federal agencies investigate AI-CSAM cases in the DFW region. Homeland Security Investigations (HSI) operates the Dallas Field Office Cyber Crimes Center and runs the largest single docket of federal CSAM investigations in the region. The North Texas Internet Crimes Against Children (ICAC) Task Force is housed at the Dallas Police Department and coordinates between federal, state, and local investigators across Collin, Dallas, Denton, and Tarrant Counties. The FBI Dallas Cyber Squad handles the federal side of investigations originating from National Center for Missing & Exploited Children (NCMEC) CyberTipline referrals, internet-service-provider reports, and undercover-operation tips. State-court prosecutions are handled by the District Attorney's Offices in Collin County (McKinney), Dallas County (Dallas), Denton County (Denton), and Tarrant County (Fort Worth), each with specialized sex-offense and crimes-against-children units.

How much does AI-CSAM defense cost, and how long does a case take?

Legal fees for an AI-CSAM defense typically run $50,000-$200,000 depending on whether the case is state or federal, the volume of imagery at issue, the available defense theories, and trial readiness. Federal cases run at the upper end because of the density of motion practice (suppression, First Amendment as-applied, Daubert), the complexity of the Sentencing Guidelines calculation, and the lifetime registration consequences. State-court § 43.26 cases run somewhat lower but require equivalent forensic-imagery expert work. Forensic-imagery and AI-detection expert costs add substantially — typically $15,000-$50,000 for retained experts. Cases typically take 18-36 months from indictment to disposition for contested cases; cases that resolve at plea can disposition in 6-12 months. Pre-arrest engagement is highest-leverage and may permit pre-indictment negotiations that change the entire case trajectory.

References

All citations link to statutes.capitol.texas.gov for primary text. Footnote numbers in the body link here; the arrow returns to the citing paragraph.

  1. Tex. Penal Code § 38.04 — Evading arrest or detention.
  2. Tex. Penal Code § 12.21 — Class A misdemeanor punishment range.
  3. Tex. Penal Code § 12.34 — Third-degree felony punishment range.
  4. Tex. Penal Code § 12.33 — Second-degree felony punishment range.
  5. Tex. Penal Code § 9.22 — Necessity affirmative defense.
  6. Tex. Code Crim. Proc. art. 38.23 — Suppression of evidence from unlawful search/detention.
  7. Tex. Code Crim. Proc. art. 39.14 — Michael Morton Act discovery.
  8. Tex. Code Crim. Proc. art. 42A.054 — 3g offenses (not including evading).
40+
Years
Combined defense experience
$0
Consult
Free initial consultation
24/7
Available
Direct-to-attorney for jail release
About the authors

The attorneys behind this page

Reggie London

Reggie London

Co-Founding Partner · Criminal Defense Attorney

Admitted in Texas, TXND, TXED, and the U.S. Court of Appeals for the Fifth Circuit. Practice spans DWI, drug, weapons, theft, and process crimes — plus federal practice.

Njeri London

Njeri London

Co-Founding Partner · Criminal Defense Attorney

Texas-licensed criminal defense attorney with deep Fourth Amendment motion practice. Focus: suppression hearings, drug-crime defense, federal-practice support.

Free Consultation · 24/7

Talk to an attorney — not a screener.

Tell us about your case. Most clients hear back within an hour. Often within minutes.

5899 Preston Rd, Ste 101 · Frisco, TX 75034

By submitting, you agree to our Privacy Policy.

Call (972) 370-5060

Attorney Advertising

This website is for general information purposes only and constitutes attorney advertising under the Texas Disciplinary Rules of Professional Conduct. Nothing on this site should be taken as legal advice for any individual case or situation. Receipt or viewing does not create an attorney–client relationship.

Past results do not guarantee similar outcomes. Each case is unique and must be evaluated on its own facts and circumstances.

L and L Law Group, PLLC attorneys are licensed to practice in the State of Texas. Njeri London (Texas Bar No. 24043266) and Reggie London (Texas Bar No. 24043514) are the attorneys responsible for the content of this site. None of the attorneys at L and L Law Group, PLLC are Board Certified by the Texas Board of Legal Specialization unless specifically and separately stated.

Please do not transmit any confidential information to L and L Law Group, PLLC by email, web form, or telephone before a written engagement is in place. Privacy Policy.

Service Areas

L&L Law Group represents clients across North Texas counties for DWI, assault, drug crimes, juvenile defense, outstanding warrants, bond reduction, and expunction matters.

Call Email Map Top
developed by MPR Digital Legal Services