Wisconsin Clarifies the “Instrument or Agent” Test: ESP CSAM Scanning Remains a Private Search Under a Totality-of-the-Circumstances Analysis

Introduction

State v. Andreas W. Rauch Sharak (2026 WI 4) presented the Supreme Court of Wisconsin with a recurring Fourth Amendment problem at the intersection of digital privacy and child sexual abuse material (CSAM) enforcement: when an electronic service provider (ESP) proactively scans user content, flags apparent CSAM, reviews it, and reports it to the National Center for Missing & Exploited Children (NCMEC), does that conduct become a government search?

Google’s automated detection flagged four files in a Google Photos account. A Google employee opened and viewed the files, after which Google submitted a CyberTip to NCMEC. NCMEC forwarded the tip to law enforcement; an officer viewed the files without a warrant, then obtained a warrant for a home/device search. The defendant, Andreas Rauch Sharak, was convicted of possession of child pornography and sought suppression, arguing Google acted as “an instrument or agent of the government,” rendering the initial viewing and the subsequent police viewing unconstitutional.

The court of appeals certified three questions, principally concerning (1) expectation of privacy in light of terms of service, (2) whether Google’s actions were private or governmental under Wisconsin law, and (3) whether police needed a warrant before viewing files received via NCMEC. The Supreme Court resolved the case primarily on the second question: Google was a private actor, and the subsequent police viewing fit within the private search doctrine.

Summary of the Opinion

  • Google was a private actor. The Court held that Google did not act as an instrument or agent of the government when it scanned for CSAM and when an employee opened and viewed the flagged files.
  • Totality-of-the-circumstances governs. The Court clarified that the considerations discussed in State v. Payano-Roman, 2006 WI 47 are not rigid “requirements,” and no single factor is dispositive; the inquiry remains the totality-of-the-circumstances test derived from Skinner v. Ry. Lab. Execs.' Ass'n, 489 U.S. 602 (1989).
  • Police could view the CyberTip files without a warrant on these facts. Because a Google employee had already viewed the images, the officer’s subsequent viewing did not exceed the private search’s scope and therefore was not a Fourth Amendment “search” under United States v. Jacobsen, 466 U.S. 109 (1984).
  • Expectation of privacy question not reached. Given the dispositive private-actor holding, the Court declined to decide whether Google’s terms of service affected a user’s reasonable expectation of privacy.

Analysis

1) Precedents Cited

Foundational “Government Participation” Framework

  • Skinner v. Ry. Lab. Execs.' Ass'n, 489 U.S. 602 (1989)
    Role in this case: Skinner supplies the governing principle that the Fourth Amendment generally restrains government, not private parties, but applies when a private party acts as “an instrument or agent of the Government.” The Court leaned heavily on Skinner’s instruction that the question “turns on the degree of the Government’s participation” and must be evaluated “in light of all the circumstances.”
    How it influenced the result: By contrasting Skinner’s intensive regulatory compulsion (rules preempting state law, restricting employee refusal, limiting employer discretion) with the CSAM reporting/moderation framework, the Court found no comparable governmental coercion or participation that would convert Google’s actions into state action.
  • State v. Payano-Roman, 2006 WI 47, 290 Wis. 2d 380, 714 N.W.2d 548
    Role in this case: Payano-Roman is Wisconsin’s core decision on private vs. governmental searches. The Court reaffirmed Payano-Roman’s adoption of Skinner’s totality approach and clarified that Payano-Roman’s listed considerations are not rigid prerequisites.
    How it influenced the result: Payano-Roman provided both the standard of review (mixed question; deference to factfinding, independent constitutional conclusion) and the analytical structure. The Court distinguished Payano-Roman’s “joint endeavor” facts (custody, officer presence, officer assistance) from Google’s independent scanning and review, undertaken before any law-enforcement involvement.
  • State v. Rogers, 148 Wis. 2d 243, 246, 435 N.W.2d 275 (Ct. App. 1988)
    Role in this case: Quoted in Payano-Roman as articulating three “requirements” for a private search.
    How it influenced the result: The Court explicitly rejected a rigid reading that would treat Rogers/Payano-Roman factors as mandatory elements. Instead, those considerations remain helpful signposts within the broader totality inquiry.

The Private Search Doctrine and Scope

  • United States v. Jacobsen, 466 U.S. 109, 115 (1984)
    Role in this case: Jacobsen is the canonical private search doctrine case: when a private party has already searched, government repetition within the same scope is not a Fourth Amendment “search.”
    How it influenced the result: Once the Court deemed Google a private actor, Jacobsen disposed of the warrantless police viewing on the record presented: a Google employee viewed the files first, and the defendant conceded police did not exceed that scope.
  • State v. Gasper, 2026 WI 3, 419 Wis. 2d 12, __ N.W.3d __
    Role in this case: Cited as a “sister case” addressing scope expansion where an ESP scanned but did not open/view particular files.
    How it influenced the result: Rauch Sharak’s case is positioned as the “employee-viewed-first” scenario, leaving more difficult “scope” issues to Gasper. The citation signals that future Wisconsin litigation will pivot on whether law enforcement exceeded the private search’s scope when ESP review is automated or incomplete.

Comparative “Government Agent” Fact Patterns

  • United States v. Booker, 728 F.3d 535, 540-41 (6th Cir. 2013)
    Role in this case: Illustrates when medical personnel can be treated as government agents: police custody, police purpose, and known/expected medical retrieval of contraband.
    How it influenced the result: Served as a contrast to show what “government participation” looks like. Google’s scan/review occurred without custody, without police presence, and without police direction.

Federal Circuit and State Supreme Court Consensus on ESP CSAM Scanning

  • United States v. Ringland, 966 F.3d 731 (8th Cir. 2020) and United States v. Miller, 982 F.3d 412 (6th Cir. 2020)
    Role in this case: Both are used to substantiate that Google/ESPs act for independent business reasons (removing abuse; protecting platform integrity) rather than as government instruments.
    How they influenced the result: They underwrote the Court’s conclusion that Google’s conduct served its own ends, and they weakened the claim that NCMEC-provided information “participated” in the search (Miller’s discussion of scanning not necessarily relying on NCMEC-provided hash values).
  • United States v. Rosenschein, 136 F.4th 1247, 1256 (10th Cir. 2025); United States v. Sykes, 65 F.4th 867, 877 (6th Cir. 2023); United States v. Rosenow, 50 F.4th 715, 730 (9th Cir. 2022); United States v. Meals, 21 F.4th 903, 907 (5th Cir. 2021)
    Role in this case: Cited for the proposition that 18 U.S.C. § 2258A’s reporting duty does not transform ESPs into government agents, emphasizing the statutory disclaimer that providers are not required to “affirmatively search, screen, or scan.”
    How they influenced the result: They collectively reinforced the Court’s view that statutory reporting obligations—without a scanning mandate—do not equate to governmental participation under Skinner.
  • Children's Health Def. v. Meta Platforms, Inc., 112 F.4th 742, 762 (9th Cir. 2024), cert. denied, 145 S. Ct. 2846 (2025)
    Role in this case: Used to characterize 47 U.S.C. § 230(c) as “entirely passive” and unlike Skinner’s coercive regime.
    How it influenced the result: Helped the Court reject the theory that § 230’s immunity structure, even if encouraging moderation, constitutes the degree of government compulsion/participation required for state action.
  • Does 1-6 v. Reddit, Inc., 51 F.4th 1137, 1145 (9th Cir. 2022)
    Role in this case: Cited to narrow the scope of § 230(e)(5)(A)’s immunity exception; liability is not triggered merely by failure to remove content absent the statute’s mens rea and conduct requirements.
    How it influenced the result: Undercut the defendant’s claim that ESPs face criminal exposure for not proactively removing trafficking-related content in a way that would coerce scanning and thereby create governmental agency.
  • United States v. Bebris, 4 F.4th 551, 560-62 (7th Cir. 2021); United States v. Richardson, 607 F.3d 357, 364-67 (4th Cir. 2010); State v. Pauli, 979 N.W.2d 39, 51-52 (Minn. 2022); State v. Lizotte, 197 A.3d 362, ¶¶22-23 (Vt. 2018)
    Role in this case: Cited for the broad consensus that ESP CSAM detection and reporting is private action.
    How they influenced the result: While not dispositive, the unanimity of authority bolstered the Court’s confidence that its analysis aligns with the prevailing constitutional treatment of these platform practices.

Wisconsin Case Law on Argument Development and Statutory Interpretation

  • Serv. Emps. Int'l Union, Loc. 1 v. Vos, 2020 WI 67, ¶24, 393 Wis. 2d 38, 946 N.W.2d 35, overruled on other grounds by, Evers v. Marklein, 2025 WI 36, 417 Wis. 2d 453, 22 N.W.3d 789
    Role in this case: Used to justify treating the case as purely Fourth Amendment (federal) because the defendant did not develop an independent argument under WIS. CONST. ART. I, § 11.
    How it influenced the result: It constrained the Court’s analysis to federal Fourth Amendment doctrine and avoided any potentially broader state-constitutional ruling.
  • Serv. Emps. Int'l Union Healthcare Wis. v. WERC, 2025 WI 29, ¶5, 416 Wis. 2d 688, 22 N.W.3d 876
    Role in this case: Cited to characterize the circuit court’s statutory-history observations as legal conclusions, not findings of fact.
    How it influenced the result: Allowed the Supreme Court to decline “adopting” the circuit court’s interpretations of 47 U.S.C. § 230 and related statutes, and to instead independently assess whether those statutes create governmental agency.

2) Legal Reasoning

a) The Court’s Clarification: Payano-Roman Factors Are Not Mandatory Elements

The Opinion’s most explicitly “new” doctrinal clarification is methodological: Wisconsin courts must not treat the three Rogers/Payano-Roman considerations as rigid “requirements” such that failure of any one factor automatically resolves the question. Instead:

  • The inquiry is a totality-of-the-circumstances assessment (Skinner; Payano-Roman).
  • Considerations like police initiation/encouragement/participation, the private party’s independent purpose, and “joint endeavor” framing are analytical guideposts, not a checklist.
  • The “ultimate inquiry” is whether the private party acted as an instrument or agent of the government given the degree of governmental participation.

This matters because “instrument or agent” disputes often arise in modern contexts (platform moderation, device repair shops, employer monitoring, private security, hospitals). A rigid element-based test can misfire in mixed-motive or mixed-contact situations; the Court reaffirmed a flexible approach keyed to actual governmental involvement.

b) Application to Google: No Government Involvement in the Search Itself

On the record before it—described as “sparse” because the circuit court denied suppression without a hearing—the Court emphasized the chronology:

  1. Google scanned files and flagged potential CSAM.
  2. A Google employee opened and viewed the flagged files.
  3. Only then did Google submit a CyberTip, starting governmental involvement.

The absence of police initiation, direction, presence, or contemporaneous assistance sharply distinguished this case from Payano-Roman’s hospital “joint endeavor” and from Booker’s police-driven medical retrieval.

c) Independent Private Purpose: Platform Integrity and User Experience

The Court gave weight to Google’s articulated non-law-enforcement purpose: enforcing its policies, ensuring compliance with law, and maintaining a “positive experience.” It treated this as an ordinary business interest—consistent with Ringland and Miller—supporting private-actor status.

d) Statutory “Encouragement” Is Not Enough Without Compulsion or Participation

The defendant’s central theory was structural: federal statutes create a constellation of duties, immunities, and liabilities that effectively deputize platforms. The Court’s response separated:

  • 18 U.S.C. § 2258A (reporting): imposes a duty to report discovered apparent CSAM but explicitly states providers need not “affirmatively search, screen, or scan.” That disclaimer “blunts” the government-agent argument and aligns with multiple federal circuits.
  • 47 U.S.C. § 230(c) (immunity for moderation): may remove liability concerns for voluntary good-faith moderation, but is “passive” and does not require or meaningfully compel the specific scanning at issue—unlike the coercive regulation in Skinner.
  • 47 U.S.C. § 230(e)(5) and 18 U.S.C. § 1591 (trafficking-related exceptions): the Court found the claimed threat of liability for “not scanning” underdeveloped and inconsistent with narrowing interpretations like Does 1-6 v. Reddit, Inc.

The through-line is Skinner’s “degree of Government’s participation”: general legislative policy that aligns private incentives with law enforcement goals is not, by itself, enough to convert a private compliance program into state action.

e) Consequence for the Officer’s Viewing: Jacobsen Controls on Scope

Because Google was private and its employee had already viewed the images, law enforcement’s later viewing of the same files did not exceed the scope of the private search. Under Jacobsen, the officer’s act was not a Fourth Amendment “search,” so no warrant was required on these facts. The Court noted the defendant’s concession on scope and pointed to State v. Gasper for harder scope questions when ESPs scan but do not open/view.

f) Narrowness: Expectation of Privacy Left Undecided

The Court expressly avoided the certified question about whether terms of service diminish a user’s reasonable expectation of privacy in provider-obtained files. This preserves future litigation on whether, and to what extent, contractual disclosure/monitoring terms alter Fourth Amendment expectations in cloud accounts—particularly outside the CSAM context.

3) Impact

a) Immediate Practical Effect in Wisconsin

  • Suppression challenges premised on “Google is the government” will be difficult absent evidence of police direction, coordination, or coercion beyond the existing federal reporting/immunity framework.
  • Litigation focus shifts to “scope” (what exactly the ESP/employee viewed; what the officer viewed; whether the officer opened additional files) and to the factual development necessary to prove agency under Payano-Roman’s clarified approach.
  • Procedural signal: the Court repeatedly noted the defendant’s burden and the sparse record. Future defendants will likely seek evidentiary hearings and discovery (where available) to attempt to show law-enforcement participation or platform–government entanglement in specific cases.

b) Doctrinal Clarification Beyond CSAM Cases

The Opinion’s clarification that Payano-Roman’s considerations are not strict requirements may affect any Wisconsin case involving private actors who uncover evidence: employers, private security, hospitals, schools, and digital service providers. Courts are instructed to avoid treating any single factor (e.g., “dual purpose,” “helping law enforcement,” or “police encouragement”) as automatically dispositive and to instead evaluate the whole relationship and degree of governmental participation.

c) Interaction with Emerging “CyberTip” Scope Disputes

By emphasizing that the officer stayed within the scope of what the Google employee already viewed—and by citing State v. Gasper—the Court implicitly frames the next frontier: cases in which automated detection occurs, but human viewing does not. Those disputes will turn on how Jacobsen’s “virtual certainty” and “scope” concepts apply to modern hashing/automated detection and whether opening additional files constitutes expansion.

Complex Concepts Simplified

Fourth Amendment “search”
A government action that intrudes on a person’s reasonable expectation of privacy (or physically trespasses in certain contexts). The Fourth Amendment generally regulates government, not private parties.
Private actor vs. “instrument or agent of the government”
A private party becomes subject to Fourth Amendment limits if the government’s involvement is sufficiently significant—through direction, participation, coercion, or a joint operation. The Court evaluates the degree of governmental participation under all the circumstances.
Totality of the circumstances
No single fact decides the case. Courts consider the entire context: who initiated the search, who participated, why it was done, and how closely the private party worked with government.
Private search doctrine (United States v. Jacobsen)
If a private party already searched and found evidence, police may repeat that same search without a warrant so long as they do not go beyond what the private party already exposed.
NCMEC and “CyberTip” reporting
ESPs report apparent CSAM to NCMEC via CyberTips; NCMEC forwards the tips to law enforcement. Federal law requires reporting of discovered apparent violations, but (per the statute’s disclaimer) does not require platforms to proactively scan.
Reasonable expectation of privacy and terms of service
A separate question (not decided here): whether a user’s agreement to monitoring/scanning terms affects what privacy expectations society recognizes as reasonable against the government.

Conclusion

State v. Rauch Sharak establishes two key takeaways for Wisconsin Fourth Amendment doctrine in the digital age. First, it clarifies that Payano-Roman’s factors are not rigid requirements; courts must apply Skinner’s totality-of-the-circumstances inquiry to decide whether a private party acted as a governmental instrument or agent. Second, applying that framework, the Court held that Google’s CSAM scanning and employee review were private actions, not transformed into state action by federal reporting and immunity statutes. Consequently, law enforcement’s warrantless viewing of the same files was permissible under United States v. Jacobsen because it did not exceed the private search’s scope.

The Opinion thus aligns Wisconsin with broad appellate authority treating ESP CSAM detection as private action, while teeing up future litigation where the decisive issue will be scope (what, exactly, the provider actually viewed) and proof of governmental participation on a developed record.