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.
