Ethics and Human Rights in Virtual Reality

AA film can show a fall from a cliff; VR can make a person's legs shake at the edge of one, even though they know they're standing on a carpet. That gap between knowing and feeling is what makes VR useful for therapy and training. It's also why ethical questions that are settled for film and flat-screen games are reopened in VR.

This page asks when a virtual experience counts morally as a real one, revisits the debate about immoral acts in games, sets out what informed consent needs to cover, looks at manipulation, and applies a human-rights frame to the companies that run virtual worlds. It ends with what a developer or researcher should take from it.cf. the broader debate on virtual acts and real harm

When Is a Virtual Experience Morally Real? FoundationalKnowledge that endures for decades — core principles

MadaryMadary and Metzinger single out the illusion of embodiment, the sense that a virtual body is one's own, as the reason VR creates risks that other media don't. They point to evidence that behaviour can change inside VR and that some changes persist after the user leaves it¹. Slater and colleagues add that people can forget, after a while in a scenario, that there's a device between them and the world, and that virtual experiences can have physical, emotional and cognitive after-effects, some of them long-lasting².

Ramirez and LaBarge give the most precise account of when this matters. They argue that VR differs from text, film and games in its "perspectival fidelity": the experience is presented from the user's own point of view, through their own body. Combined with "context realism", a scenario that plausibly could happen, the result is a "virtually real experience". For those, they propose an equivalence principle: if it would be wrong to allow someone to have an experience in reality, it would be wrong to allow them to have it in a virtually real setting³.perspectival fidelity: the view is from your eyes

The principle is deliberately limited in scope. It doesn't say that everything wrong in reality is wrong in VR; a stylised zombie game played in a headset has neither the context realism nor, usually, the fidelity the principle needs. It does say that a sufficiently realistic, first-person simulation of, for example, being assaulted needs the kind of justification and safeguards that exposing someone to the real thing would need. Ramirez and LaBarge intend it as a tool for ethics committees and researchers assessing risk.

Immoral Acts in Virtual Worlds: The Gamer's Dilemma Applied / MethodologicalKnowledge with a 5–10 year half-life — stable practice

AA separate question concerns what a player does rather than what is done to them. Luck's "gamer's dilemma" starts from a widely held pair of intuitions: virtual murder in games is usually thought acceptable, and virtual paedophilia is not. Luck argues that the standard arguments offered in defence of virtual murder (no one is harmed, it's only fiction, it may even be cathartic) seem to apply equally to virtual paedophilia, so either some morally relevant difference has to be found or one of the intuitions has to go⁴. Bartel's response is that virtual paedophilia depicts sexual acts involving children and so falls under the wrongs of child sexual abuse imagery, which supplies the morally relevant difference from virtual murder⁵. The debate has continued since, without a settled resolution.

Debates about whether violent games desensitise players or offer a harmless outlet were mostly framed around flat-screen games watched through a window. Our reading of Ramirez and LaBarge is that their distinction explains why those debates don't transfer straight to VR: the "it's only a representation" defence weakens as perspectival fidelity rises, because the player is performing the act with their own body rather than directing a character. That makes the design of what VR lets players do to other characters, and especially to other people's avatars, an ethical decision as well as a creative one.the line between fiction and act blurs

What Informed Consent Has to Cover Applied / MethodologicalKnowledge with a 5–10 year half-life — stable practice

ConsentConsent in VR has two settings with different standards. In research, Madary and Metzinger's recommendations cover the limits of what experiments should do to participants, informed consent (which should state explicitly that immersive VR can have lasting behavioural influences, some of them not yet known), clinical risks, dual use, and experiments run online¹. They also caution that a code of conduct can't replace ethical reasoning by the researchers themselves. Slater and colleagues add practical safeguards: a stop button that respects a participant's right to end the experience at any time, aftercare to help people readjust when they come out, and longer-term follow-up².

For consumers, consent mostly means a terms-of-service screen. That's a poor fit for VR, where what is collected (see Privacy in VR) and what an experience may do to the user emotionally are both hard to grasp before trying it. Consent that a user can act on needs to be specific to the experience, given at the point it becomes relevant, and revocable during it: content warnings before intense scenes, a clear way to pause or leave, and separate permission for sensitive sensors such as eye tracking.biometric data is often collected by default

Realism, Persuasion and Manipulation Applied / MethodologicalKnowledge with a 5–10 year half-life — stable practice

TheThe same realism that makes VR useful makes it persuasive. Slater and colleagues discuss "superrealism", virtual humans realistic enough that harmful content becomes more convincing, and "identity hacking": a trusted person's avatar taken over by someone else, or a real person shown doing something in XR they never did². They also note that XR can separate sensory experience from what happened, which threatens the shared evidence that public argument relies on.

Add eye tracking and the risk changes in kind. A system that can see which elements of a scene hold a user's attention, and adjust the scene in response, can tune persuasion to that individual's involuntary reactions. This is sometimes called "cognitive hacking". Heller's "biometric psychography" names the data that would drive it⁶. The line a developer should hold is between adapting an experience for a user, with their knowledge (difficulty, comfort, accessibility), and adapting it to steer them towards someone else's goal, such as a purchase, a belief or more time spent, without their knowledge.the body becomes the data

A Human-Rights Frame for Virtual Worlds FoundationalKnowledge that endures for decades — core principles

Human-rightsHuman-rights law binds states, but the UN Guiding Principles on Business and Human Rights (2011) set out what is expected of companies. They rest on three pillars: the state duty to protect against human-rights abuses by third parties, including businesses; the corporate responsibility to respect human rights, which means carrying out due diligence to identify and address the harms a company's operations cause or contribute to; and access to effective remedy for those affected⁷. Applied to a VR platform, the second and third pillars translate into risk assessment before launching features, and working reporting, appeals and redress after something goes wrong.

Heller argues for building privacy and human rights into immersive technology at the design stage rather than leaving them to later regulation⁶, which matches the Guiding Principles' emphasis on due diligence before harm occurs.

Where Governance Has Gaps Ephemeral / ToolingKnowledge that evolves in months to a year — check for updates

NoNo regulator deals with immersive technology as such. Existing instruments each cover part of the ground:

Slater and colleagues call for a cross-industry code of conduct for producers of virtual content². Any such code, or any future regulation, is more likely to hold up if technologists, ethicists, policymakers and disabled and other affected users draft it together, since each group sees failures the others miss.

What This Means When You're Building It Applied / MethodologicalKnowledge with a 5–10 year half-life — stable practice

• Apply the equivalence test to intense scenes. If a first-person, realistic scenario would be unacceptable to put someone through in reality, it needs strong justification, warnings and support in VR too.
• Always provide a way out. A pause or exit that works at any moment, and a calm space to return to, are basic requirements.
• Get consent at the point it matters, separately for sensitive sensors and intense content, rather than once at install.
• Adapt for the user, never against them. Use what you learn from tracking to improve comfort and accessibility, not to steer behaviour without the user knowing.
• Think through misuse before launch. Ask how a feature could be used to harass, impersonate or exclude someone, and design for that case.
• Take user testing through ethics review. Student projects count.

Related Topics

• Virtual Reality and Games: Ethics, Law and Society: the overview of this set.
• Freedom of Expression in Virtual Spaces: Article 10 in privately run worlds.
• Privacy in VR: When the Body Becomes the Data: the data behind psychography and manipulation.
• Accessibility and Inclusion in Games and VR: non-discrimination in design practice.
• Being the God of a Virtual World: the ethics of power over a virtual world's inhabitants.
• Research Ethics and AI Ethics: the wider frameworks these questions sit in.

References