In eagerly-awaited decision, SCOW holds that police did not exceed scope of previously conducted “private search” of suspected CSAM
State v, Michael Joseph Gasper., 2026 WI 3, 1/14/26, affirming a published decision of the court of appeals (on other grounds); case activity
In a narrow decision, SCOW holds that law enforcement did not exceed the scope of a private search conducted on a file uploaded by Gasper to Snapchat account which allegedly contained CSAM and affirms.
(See our prior posts here and here).
Snapchat uses a “hash-based scanning program” known as “PhotoDNA” to detect and flag suspected child sexual abuse material (“CSAM”) hosted on its servers. (¶7). This technology analyzes the “hash value” of a given photo to determine if it corresponds with known CSAM. (Id.). Here, Snapchat flagged a single 16-second video using this technology and reported the suspected file to NCMEC. (Id.). No one at Snapchat actually viewed the file before alerting NCMEC. (Id.). “NCMEC also did not view the video, but did confirm that the video was CSAM through a hash match of the uploaded file to visually similar files that were previously viewed and categorized by NCMEC.” (¶8). It transmitted the suspected file to the DOJ, which viewed the file and confirmed it was CSAM. (Id.). After the internet service provider identified Gasper as the likely uploader of the material, another officer with local law enforcement again viewed the image to confirm it was CSAM. (¶9). It was at this point that law enforcement sought a warrant, which revealed additional CSAM in Gasper’s possession. (Id.). Gasper moved to suppress, arguing that the government’s viewing of the file without a warrant was an unconstitutional search which tainted all subsequent evidence. (¶10). The trial court granted the motion. (¶15). COA reversed, essentially holding that Gasper had given up his reasonable expectation of privacy in his Snapchat files via the terms of service he agreed to by virtue of accessing the Snapchat platform. (¶17).
All seven justices would affirm the decision of COA, although they disagree as to how and why.
Five justices (Ziegler, Karofsky, Bradley, Hagedorn and Protasiewicz) join a majority opinion holding that this case is resolvable via application of the “private search” doctrine which allows an agent of the government to “view what a private party has freely made available for his inspection without offending the Fourth Amendment.” (¶22) (cleaned up). “A private search, when repeated by the government, does not then become a government search, unless that search exceeds the scope of the private search.” (¶23). Notably, when the government alleges a private search, it becomes the defendant’s burden to prove that a government search occurred via a preponderance of the evidence. (Id.)
Here, Gasper alleges that law enforcement exceeded the scope of any private search conducted using computer software when it visually examined the suspected file. (Id.). However, SCOW ultimately relies on a “virtual certainty” test to uphold the law enforcement conduct. Under that test, “the private search doctrine has been found to be applicable when there is a ‘virtual certainty’ that the government’s search will not reveal anything more than that which the private party represented.” (¶31). SCOW therefore sides with the decisions of the Fifth and Sixth Circuits and holds that opening a file and conducting a visual inspection to confirm that the flagged file is what the private party says it is does not exceed the scope of an earlier private search of the file conducted using a computer program. (¶40). It is irrelevant to SCOW that one search uses a computer program analyzing “hash values” and the other involves a visual inspection by an actual human being. (Id.). “Whether law enforcement may glean something more from viewing the same file does not amount to an expansion of the initial search.” (¶42). The test focuses on what has been revealed and, in SCOW’s view, “The private search is not exceeded if the identical information, here a single, 16-second video, is scanned and flagged as CSAM by a private actor, then provided to law enforcement for review and they look at the video.” (Id.). SCOW therefore affirms the decision of COA.
Justice Ziegler concurs with her majority opinion in order to clarify: (1) Snapchat was following federal law when it passed along the suspected CSAM; (2) that because the Cyber Tip alone probably generated sufficient probable cause for the ensuing search warrant, it was not unreasonable for law enforcement to take the less intrusive step of viewing the file first; (3) the high bar for proving a lack of good faith in a Fourth Amendment context and (4) that Gasper’s insistence that ESPs view suspected CSAM before passing it on to law enforcement could run afoul of victims’ rights, privacy laws, and create needless secondary trauma.
Justice Dallet also concurs in the result. She writes separately in order to air her disagreements with the rationale of the court of appeals. Specifically, she finds the reasonable expectation of privacy argument, under which users like Gasper can waive Fourth Amendment rights via terms of service dictated by corporate tech overlords (our gloss), to be a concerning reading of Fourth Amendment requirements. As to the majority’s private search analysis, Justice Dallet also agrees with Justice Crawford that the police exceeded the scope of any private search by opening and viewing the file. While she believes the issue of good faith is a close call given the unsettled nature of this legal question, she ultimately finds that the good-faith exception rescues the conduct of law enforcement and would affirm COA on that basis.
Justice Hagedorn, joined by the remaining members of the majority (Karofsky, and Protasiewicz), files his own concurrence which essentially offers a different gloss on the private search analysis.
Finally, Justice Crawford, partially joined by Justice Dallet, concurs in part and dissents in part. Justice Crawford disagrees with the majority and asserts that the visual viewing of the file is constitutionally significant: “The digital scans conducted by ESPs provide only a binary determination that a file falls within a category of contraband, and even then, the classification is a tentative one: that a file is potential CSAM. Only the officers’ subsequent viewing of the video confirmed that the file contained intact CSAM and, in this case, disclosed the unambiguously illegal nature of the content under Wisconsin law.” (¶119). Justice Crawford analogizes to a tip that law enforcement may receive about a suspected drug house; even if that information is highly reliable, law enforcement still needs a warrant to enter the home and confirm the observations of the tipster. (¶120). Justice Crawford further observes that the police did not repeat the private search by conducting their own software scan; they went further and physically opened the file to visually inspect it. (¶123). Justice Crawford therefore rings the alarm that the decision “sanctions greater government intrusion in reliance on private companies’ technological tools. Condoning warrantless government searches that surpass a commercial entity’s previous intrusion into places in which citizens reasonably expect privacy undermines the Fourth Amendment’s protections against unreasonable searches.” (¶124). And, unlike Justice Dallet, Justice Crawford believes that police were suitably on notice that they should have obtained a warrant before opening the suspected video. (¶133). She would not find that good-faith excuses their conduct, although she ultimately agrees that the the motion to suppress should not have been granted. In her view, the hash-value match and the subscriber information were sufficient to generate probable cause for a search of Gasper’s residence, regardless of whether anyone actually opened and viewed the suspected CSAM. (¶137).
And with that, we have SCOW’s first major decision of 2026, a hotly-anticipated Fourth Amendment case which sees SCOW wading into a legal issue that appears destined for future SCOTUS resolution given the potential circuit split.
Ultimately, however, the case is notable for what it doesn’t decide.
First, there’s no dispute that the initial search by Snapchat was a private search and no claim that this was in any way a joint endeavor or a quasi-government search. (¶16 n.5). However, as Justice Ziegler points out in her concurrence, there is federal law which appears to enlist Snapchat in the enforcement of CSAM laws. That issue is directly presented in the companion case argued on the same day, Rauch Sharak, and a decision on that case is still forthcoming.
Second, the published decision of COA understandably generated a fair amount of concern from digital privacy advocates, as it appears to stand for the proposition that Fourth Amendment rights can be forfeited via contracts of adhesion mandated by tech companies as a condition of living life online and accessing major platforms. Wisconsin therefore saw a deluge of interesting amici, including some of these aforementioned tech giants, taking issue with precisely that reading. However, SCOW ducks that issue here, as well, although it remains viable in the pending Rauch Sharak appeal. Although Justices Dallet and Crawford appear to have unambiguously signaled their positions on that matter, it remains to be seen whether the unique majority coalition presented here will hold in that case. For now, however, the sweeping decision of COA on this point has been vacated and without precedential effect. (¶6 n.6).
Third, as Justice Dallet points out in her concurrence, the issue of good faith is complex given the unsettled nature of this legal question. SCOW’s precedent is notable inasmuch as it employs a more stringent good faith standard than that required by SCOTUS. Justices Crawford, Ziegler, and Dallet appear to have staked out a wide spectrum of positions on the issue and their separate writings will be useful for readers litigating the good faith issue.
Fourth, the decision also does not resolve the impact of reliability challenges as to software tools like PhotoDNA in applying the “virtual certainty” test. At one point, for example, SCOW suggests that reliability may be irrelevant: “Even if PhotoDNA was deemed unreliable, which we need not decide, it is less than clear how that would impact the analysis of the private search doctrine since the analysis focuses on what the private actor searched.” (¶36). Elsewhere, however, SCOW acknowledges that challenges to the reliability of a software identification may in fact impact the “virtual certainty” analysis. (¶32 n. 18). SCOW’s mixed signals on this point may be fodder for future litigation.
Finally, SCOW also doesn’t resolve another issue implicitly lurking within this case: whether the visual inspection matters or whether, as both Justices Ziegler and Crawford appear to agree, a mere software-generated tip can generate probable cause.
Setting aside what SCOW did not resolve, this remains an an interesting Fourth Amendment issue, one that has divided federal courts. As Justice Ziegler observes, the Fifth and Sixth Circuits are on Wisconsin’s side. However, the Second and Ninth have reached an opposite conclusion. While a reader could get bogged down in technological details and doctrinal disagreements, the ultimate takeaway from this case is that SCOW does not believe opening a file to confirm it is, in fact, CSAM exceeds the scope of an earlier private search conducted using sophisticated computer software. In SCOW’s view, visual inspection essentially accesses the same “information” already revealed via the private search (even if, as Justice Crawford points out, the two searches are actually “looking” at totally different things). Although the justices all offer preferred analogies, it is this relatively straightforward understanding of the issue that carries the day for the State.
It is unclear, at this time, whether this understanding of the “virtual certainty” test will apply only in context of such digital evidence cases or whether, as the separate writings indicate, SCOW’s majority opinion signals a subtle shift in the complex arena of so-called “private” searches. Stay tuned for future developments.