Last updated: 2026-11-18

U
Undergraduate level

Crime in Games and Virtual Worlds

Players have been robbing each other in online games for twenty years, and police have been arresting some of them for almost as long. What has changed is less the behaviour than the stakes: game items now sell for real money, social VR puts a stranger's avatar within arm's reach of a user's tracked body, and the platforms hosting all of this run across dozens of legal systems at once.

This page sorts virtual-world crime into three kinds, follows two of them through the courts, and looks at why investigating any of it is hard. It ends with what a developer can do about each.

Three Kinds of Virtual-World Crime FoundationalKnowledge that endures for decades โ€” core principles

INTERPOL's 2024 white paper on the metaverse lists the offences its expert group expects to see in virtual worlds, from grooming and radicalisation to theft of virtual property, trespass into private virtual spaces and "robbery from an avatar"1. A UCL expert study published the same year identified 30 distinct metaverse crime threats and found that both of its expert panels rated sexual offences and crimes against the person, such as hate crime, as the most harmful and the hardest to address2.

A long list like that becomes easier to reason about once it is grouped by what is being attacked, because each group meets the law in a different place:

TargetExamplesWhere existing law reaches
The systemAccount takeover, malware on a headset, cheating tools that tamper with serversWell covered: unauthorised access is an offence whether or not a game is involved
Things of valueStolen items or currency, phishing for game logins, scams in item tradingFraud is well covered; whether a game item can be stolen depends on the country
The personHarassment, stalking, grooming, sexual assault on an avatarCommunications offences reach some of it; offences defined by physical contact mostly don't

Attacks on the system are the least new. Giaretta's survey of VR security separates threats peculiar to VR, such as moving the boundary of a user's play area so they walk into a real wall, or inserting unwanted objects into the scene, from generic threats that apply to any networked device3. In the UK, either kind falls under section 1 of the Computer Misuse Act 1990, which makes it an offence to cause a computer to perform any function with intent to secure unauthorised access to a program or data4. A headset is a computer. The two rows below it are where the law runs into difficulty.

graph TD A["Something happens
in a virtual world"] --> Q1{"Was a system
accessed without
authorisation?"} Q1 -->|Yes| CMA["Computer misuse
offence"] Q1 -->|No| Q2{"Was someone
deceived for gain?"} Q2 -->|Yes| F["Fraud offence"] Q2 -->|No| Q3{"Was an item
taken?"} Q3 -->|Yes| T["Theft, if the item
counts as property
(varies by country)"] Q3 -->|No| Q4{"Was a person
targeted?"} Q4 -->|Yes| P["Communications or
harassment offences;
contact offences
usually out of reach"] Q4 -->|No| G["Game rules and
platform moderation"]

Read the chart as a sketch of where to start looking. A single incident often lands in more than one box, as the Habbo case below did.

When a Virtual Item Counts as Property Applied / MethodologicalKnowledge with a 5โ€“10 year half-life โ€” stable practice

Three cases from the 2000s and early 2010s show the law moving on this question.

In 2005, police in Kagawa, Japan, arrested a Chinese student who had used automated characters (bots) in Lineage II to beat other players' characters and take their items, which he then sold for real money. Contemporary reporting noted that Japan had no law treating virtual items as property, so the realistic charge was unauthorised access rather than theft5.

Two years later, Dutch police arrested a 17-year-old accused of taking about โ‚ฌ4,000 worth of virtual furniture from other users' rooms in Habbo Hotel. The furniture had been bought with Habbo credits, which cost real money, and reports said the suspects obtained victims' passwords through fake Habbo websites6. Phishing a password and logging in as someone else is fraud and computer misuse in most countries, so this case never had to settle whether the furniture itself could be stolen.

The Dutch Supreme Court (Hoge Raad) settled it in 2012. Two youths had beaten and threatened a 13-year-old with a knife until he logged into his RuneScape account and transferred a virtual amulet and mask to one of them. The court held that the items were "goods" under Article 310 of the Dutch Criminal Code and could be the object of theft, because they had value to the people involved and the victim had exclusive, factual control over them7.

English law has no equivalent ruling. The Theft Act 1968 defines property broadly, as money and all other property "including things in action and other intangible property"8, but most game terms of service say the publisher owns every item and the player holds only a licence to use it. Our reading is that a prosecutor here would reach first for the Fraud Act 2006, whose offence of fraud by false representation needs no finding about who owns the item9, or for the Computer Misuse Act.

When the Harm Is to the Person Applied / MethodologicalKnowledge with a 5โ€“10 year half-life โ€” stable practice

Orin Kerr argued in 2008 that criminal law looks at physical conduct, not at what participants experience inside a game. A virtual act should become a crime, on his account, only when its harm leaves the game, and legislatures should be slow to create new offences for harms that stay inside it10. That view was written when most virtual worlds were seen on a monitor. A headset changes the experience it was written about.

In January 2024 it was reported that UK police were investigating an attack on the avatar of a girl under 16 by several adult users' avatars in a VR game. Officers said she had no physical injury but had suffered psychological trauma comparable to rape, and the then Home Secretary, James Cleverly, said that while such incidents are easy to dismiss as "not real", the whole point of these environments is that they are "incredibly immersive"11. The difficulty for any prosecution is in the statute. Under section 3 of the Sexual Offences Act 2003, sexual assault requires that a person "intentionally touches another person", and section 79(8) defines touching as contact with the body, with anything else, or through anything12. An avatar's hand passing through another avatar meets none of those.

Some of what happens in such incidents is reachable by other routes. Threatening or grossly offensive messages, stalking and grooming are all offences however they are delivered, and Lemley and Volokh discuss whether civil wrongs such as intrusion upon seclusion could cover conduct like virtual groping13. The physical-contact offences remain the gap, and haptic suits that let a user feel a touch will press on it further. Harassment as a design and moderation problem is covered on the freedom of expression page; the moral question of whether a virtual act can wrong someone is on the ethics and human rights page.

Why Investigation Is Hard Ephemeral / ToolingKnowledge that evolves in months to a year โ€” check for updates

Even where an offence clearly exists, three problems slow every case.

  • Jurisdiction. The victim, the suspect, the platform company and its servers can sit in four different countries. INTERPOL's white paper names this, along with the lack of common standards between platforms, as a main obstacle for investigators1.
  • Evidence that changes. A virtual room can be edited, reset or deleted, and a user's view of an incident exists only on their device unless someone recorded it. What the platform logged, and for how long it keeps the logs, decides whether there is a case.
  • Attribution. An avatar name leads to an account, an account to an email address, and the email address may lead nowhere. Linking any of it to a person usually needs a disclosure request to the platform, which is where the next point comes in.

Cross-border requests rely on treaties. The Council of Europe's Budapest Convention on Cybercrime (2001), to which the UK is a party, harmonises core offences and sets up procedures for preserving and sharing evidence between member states14; several large states, including Russia and China, have never joined it. A broader United Nations Convention against Cybercrime was adopted by the General Assembly in December 2024 and opened for signature in Hanoi in October 2025, where the UK signed15. It takes effect only after 40 states ratify it.

The data that would identify a suspect is often body data, such as motion traces and voice, which brings the investigation up against the privacy questions on the privacy and body data page.

VR as a Tool for Investigators Applied / MethodologicalKnowledge with a 5โ€“10 year half-life โ€” stable practice

The same technology is also used to train and research. Choi and colleagues piloted a VR program that taught students the search-and-seizure procedure for digital evidence; pre- and post-tests and interviews suggested it helped them understand and reinforce the concepts and steps16. In criminology, van Gelder, Otte and Luciano argue that VR lets researchers put participants into realistic scenarios, such as a decision to commit a crime, that would be unethical or impossible to stage for real, while keeping the experimental control of a laboratory17.

What This Means When You're Building It Applied / MethodologicalKnowledge with a 5โ€“10 year half-life โ€” stable practice

  • Protect accounts first. Almost every property case above began with someone else logging in. Offer two-factor authentication and warn about lookalike login pages.
  • Design trades to be reversible. A short hold on high-value transfers, or a confirmation step the victim can refuse, makes a coerced or fraudulent transfer harder to complete and easier to undo.
  • Decide what to log for evidence, and for how long. A report button that saves the last few minutes of positions and voice from the reporter's point of view gives moderators and police something to work with. Recording everyone all the time would be a privacy failure, so keep the window short and tie it to a report.
  • Make reporting fast and in-world. A target who has to take off the headset to find a web form will often not bother.
  • Know which law you're under. If your players are in the UK, the Online Safety Act's duties for user-to-user services apply to a multiplayer game as much as to a social network.

References


  1. INTERPOL. (2024). Metaverse: A law enforcement perspective [White paper]. https://www.interpol.int/content/download/20828/file/Metaverse%20-%20a%20law%20enforcement%20perspective.pdf โ†ฉโ†ฉ

  2. Gรณmez-Quintero, J., Johnson, S. D., Borrion, H., & Lundrigan, S. (2024). A scoping study of crime facilitated by the metaverse. Futures, 157, 103338. https://doi.org/10.1016/j.futures.2024.103338 โ†ฉ

  3. Giaretta, A. (2025). Security and privacy in virtual reality: A literature survey. Virtual Reality, 29, Article 10. https://doi.org/10.1007/s10055-024-01079-9 โ†ฉ

  4. Computer Misuse Act 1990, s.1. https://www.legislation.gov.uk/ukpga/1990/18/section/1 โ†ฉ

  5. The Register. (2005, August 19). Man mugs online game characters. https://www.theregister.com/2005/08/19/online_mugging/ โ†ฉ

  6. Engadget. (2007, November 14). Dutch teen arrested for stealing virtual furniture. https://www.engadget.com/2007-11-14-dutch-teen-arrested-for-stealing-virtual-furniture.html โ†ฉ

  7. Hoge Raad (Supreme Court of the Netherlands), 31 January 2012, ECLI:NL:HR:2012:BQ9251 (RuneScape). https://uitspraken.rechtspraak.nl/details?id=ECLI:NL:HR:2012:BQ9251 โ†ฉ

  8. Theft Act 1968, s.4(1). https://www.legislation.gov.uk/ukpga/1968/60/section/4 โ†ฉ

  9. Fraud Act 2006, s.2. https://www.legislation.gov.uk/ukpga/2006/35/section/2 โ†ฉ

  10. Kerr, O. S. (2008). Criminal law in virtual worlds. University of Chicago Legal Forum, 2008, 415โ€“429. https://papers.ssrn.com/abstract=1097392 โ†ฉ

  11. Marsden, H. (2024, January 2). Rape in the metaverse: A case for the real-life police? The Week. https://theweek.com/crime/rape-metaverse-real-life-police-crime โ†ฉ

  12. Sexual Offences Act 2003, ss.3 and 79(8). https://www.legislation.gov.uk/ukpga/2003/42/section/3 โ†ฉ

  13. Lemley, M. A., & Volokh, E. (2018). Law, virtual reality, and augmented reality. University of Pennsylvania Law Review, 166(5), 1051. https://papers.ssrn.com/abstract=2933867 โ†ฉ

  14. Council of Europe. (2001). Convention on Cybercrime (ETS No. 185). https://www.coe.int/en/web/cybercrime/the-budapest-convention โ†ฉ

  15. UK Government. (2025). UK national statement on signing the UN Convention against Cybercrime. https://www.gov.uk/government/speeches/uk-national-statement-on-signing-the-un-convention-against-cybercrime โ†ฉ

  16. Choi, K.-S., Lee, H., Park, G., & Han, C. (2022). Virtual reality program in cybercrime investigation: A pilot study examining search and seizure of digital evidence practice. Cyberpsychology, Behavior, and Social Networking, 25(1), 43โ€“50. https://doi.org/10.1089/cyber.2020.0894 โ†ฉ

  17. van Gelder, J.-L., Otte, M., & Luciano, E. C. (2014). Using virtual reality in criminological research. Crime Science, 3, Article 10. https://doi.org/10.1186/s40163-014-0010-5 โ†ฉ