Do We Possess Our Cache Files?

The last two decades have forced criminal law into unfamiliar terrain. Our devices record and retain data in ways we do not see, let alone control. Among the most contested examples are cache files: temporary copies of internet content that browsers store automatically to speed up browsing. They sit buried in directories that most users will never open, let alone understand. Yet when the underlying material is illicit—terrorism manuals, propaganda videos, or images of child sexual abuse—prosecutors have been tempted to treat the mere presence of cached files as criminal “possession”.

The result is a legal puzzle that exposes the fault line between human agency and machine automation. At stake are fundamental questions of mens rea: when can the law punish someone for having digital material they never consciously chose to store? The issue is not academic. Convictions for possessing child exploitation images or terrorism-related documents can bring years in prison, and sometimes it is only the hidden cache that the prosecution can point to.

This essay explores how courts across common law jurisdictions have approached cached files. It focuses on the requirement of knowledge and control as essential elements of possession, with particular emphasis on the Ninth Circuit’s reasoning in Kuchinski¹. It then considers how technical realities and evolving doctrines of mens rea should guide a principled way forward.

Mens Rea and the Texture of Possession

Possession offences have always required a mental element. Anglo-American law distinguishes between actual possession (physical custody of an item) and constructive possession (the ability and intent to control it). Both categories presuppose that the defendant knows what they have. A man who unknowingly carries a package of heroin planted in his bag does not “possess” it in law.

Digital cases strain this traditional framework. The automatic nature of caching means that material can exist on a device without the user’s awareness. The cache is not like a file placed on the desktop or stored in a “downloads” folder. It is more akin to dust caught by the filters of a machine: technically present, but invisible unless one knows where to look.

The German tradition usefully breaks down intention into Wissen (knowledge) and Wollen (will or desire). Both must be present for culpability. If a user has neither awareness of the cache nor any wish to retain its contents, treating those files as possessed collapses the distinction between voluntary conduct and happenstance.

The danger is obvious: if possession is defined as whatever a device happens to store, then criminal liability becomes strict. Every act of online browsing would carry the risk of unwitting criminality.

The Kuchinski Case

The Ninth Circuit’s decision in Kuchinski² remains the leading American authority on this point. FBI agents had discovered between 15,000 and 19,000 images of child pornography on the defendant’s computer. Of these, a small number were in his “downloads” folder or the recycle bin—files he had actively saved or at least handled. But the vast majority, over 17,000 images, resided in temporary internet cache files.

The government sought to treat every cached image as part of the sentencing calculation, arguing that their presence showed possession. The Ninth Circuit rejected that approach. Unless the prosecution could show that the defendant knew about the cache and could access or control it, those files could not count toward possession. The court stressed:

This reasoning captures the essence of the mens rea requirement. Mere technical presence of data is not enough; what matters is whether the defendant could realistically exercise dominion over it. The ruling effectively drew a line between images deliberately saved and those passively cached.

Kuchinski has not gone unchallenged. Other circuits have adopted broader views, inferring knowledge from a pattern of internet use. The Fifth and Eleventh Circuits, for example, have reasoned that repeated access to prohibited material implies awareness that the computer retains some record of it. Yet this inference risks collapsing the distinction between viewing and possessing, criminalizing the very act of clicking on a webpage. The Tenth Circuit has required proof that a defendant actually knew of and could access cache files (Tucker).³ Between these poles sits the Ninth Circuit’s later decision in Romm,⁴ which emphasized capacity for control, but again raised questions about equating theoretical capability with practical possession.

Cache Files as a Technical Reality

Understanding why the Ninth Circuit’s approach makes sense requires a short detour into technology. Browsers cache content automatically to improve speed and reduce bandwidth. The user does not choose which images or scripts are stored; the browser decides according to algorithms. Cache directories are hidden from ordinary navigation and can only be explored with technical knowledge. Files may disappear within hours or linger for weeks, depending on settings no ordinary user monitors.

This is not equivalent to saving a file to the desktop. The latter reflects an act of will, the former a machine’s background process. To treat both as possession is to ignore the element of human agency that underpins criminal responsibility.

The volatility of cache data also matters. Some jurisdictions have distinguished between files stored in random access memory (RAM), which vanish when the power goes off, and those written to a hard drive. But even hard-drive caches are unstable; they overwrite themselves as new material is viewed. Treating such transient storage as equivalent to keeping contraband in a locked drawer risks exaggerating the permanence and significance of what are, in effect, fleeting technical residues.

Comparative Judicial Approaches

Courts in other common law jurisdictions have likewise struggled. In Cyprus, the case of Police v AM (17985/2015) invoked Atkins v DPP, holding that automatic caching from online viewing does not constitute possession unless the defendant knew the material existed on the device. The court recognized that knowledge of having seen content online is not the same as knowledge of possessing a local copy.

English law faces similar tensions under Section 58 of the Terrorism Act 2000, which criminalizes possession of information “likely to be useful” to terrorists. If browsers automatically cache open-source material—bus timetables, technical manuals—does a user possess it? The statute includes a “reasonable excuse” defence, but whether that stretches to cache files remains unsettled.

American courts beyond the Ninth Circuit have fractured. The Tenth Circuit demands direct proof of knowledge and access (Tucker)⁵; the Fifth and Eleventh Circuits infer knowledge more readily. Romm⁶ tried to bridge the gap by focusing on control, but Kuchinski reminds us that practical ability, not theoretical possibility, should be the standard.

Mens Rea Doctrines in Flux

Academic debate about mens rea in the digital context helps illuminate the stakes. JJ Child and Adrian Hunt argue that the law still clings to a “present-fault paradigm“, assuming that mental states map neatly onto present conduct. ⁷ Cache cases show how inadequate that paradigm can be: the user’s act is browsing, while the supposed possession is a collateral effect generated by software.

Similarly, Alexander Sarch’s exploration of “collective knowledge” highlights the risks of imputing awareness that no single person actually has.⁸ Translating that into the digital sphere, imputing knowledge of caches to the average user risks fictionalising awareness simply because prosecutors or forensic experts know where to look.

And as Findlay Stark has warned, courts sometimes indulge in “recycled malice“, projecting intent across contexts simply because it feels morally fitting.⁹ With cache files, the danger is that culpability for seeking illicit websites is casually extended to possession of hidden technical copies.

Terrorism, Digital Ecology, and the Expanding Net

The issue is particularly acute in terrorism cases. Scholars studying the “digitalised terrorism ecology” have shown how extremist networks thrive within open online infrastructures.¹⁰ Governments respond with expansive possession offences, but when these collide with caching, the net of liability is cast too wide.

Federica D’Alessandra and Ross Gildea argue that rapid technological change requires atrocity-crime analysis to integrate digital infrastructures, a lesson equally applicable to domestic terrorism statutes.¹¹

The Evidentiary Dimension

From a forensic perspective, cached material also raises evidentiary problems. As Christos Karagiannis and Kostas Vergidis explain, digital evidence is fragile and context-dependent, requiring authenticity and reliability before it can ground liability¹². For caches, that means courts should be wary of treating automated storage as evidence of knowing possession.

Conclusion: Human Agency in a Machine World

Cache files epitomize the modern challenge to criminal law. They are the digital residue of ordinary browsing, created without will or awareness. To call them “possessed” risks imposing liability for acts the human never performed.

The Ninth Circuit’s decision in Kuchinski, reinforced by the principles in Tucker and Romm, remains the most principled guide. Otherwise, the law risks drifting into a regime of strict liability for the unconscious by-products of technology.

The larger question is whether criminal law will remain anchored in human agency. As digital systems proliferate—artificial intelligence, predictive algorithms, automated storage—the temptation to criminalize technical presence will only grow. Cache files are the test case. Insisting that mens rea requires genuine awareness, not just machine residue, is essential if the criminal law is to retain its moral legitimacy in the digital age.