Cybersecurity and the Metaverse

While much has been written about the opportunities of the metaverse, less has been written about the cybersecurity risks that come with it. Malicious actors can exploit vulnerabilities in XR hardware and software to compromise the confidentiality, integrity, and accessibility of the data or functions of an XR program. Compromising an XR technology will be just another means for bad actors to perform ransomware attacks, steal intellectual property, or disrupt National Critical Functions.¹ National security and cybersecurity practitioners must also determine whether a new novel cyber attack could be launched through XR technologies. And this is not to mention other security concerns that will stem from the use of XR technologies, such as the spread of mis-, dis-, and mal-information, theft of virtual currencies, online radicalization challenges, and data privacy concerns. While this section will focus on how bad actors can impact the confidentiality, integrity, and accessibility of XR and associated technologies and touch on data privacy concerns, national security practitioners must also be cognizant of other security concerns that will surely arise with the use of metaverses.

Today’s hardware devices, such as a headset or someone’s smartphone, tablet, or laptop, are what allow users to access a metaverse. While much is understood about the security vulnerabilities of the latter, the increasing proliferation of headsets adds new capabilities and vulnerabilities that previous devices were limited in. For instance, headsets can capture the user’s vocals, audio sounds in the room, retinal movements, user location, orientation, body movements, and a 3D mapping of the user’s environment.² Peeking inside the headset reveals the same underlying supply chain issues that have plagued other information and communications technologies, particularly semiconductors. The same chips powering our phones, computers, cars, and other interactive technologies are the same ones found in these headsets. Policymakers have implemented a slew of policies to address the semiconductor supply chain, but there may be a slew of new supply chain issues with headsets, such as the near-eye displays on the headset. China, for instance, has detailed in a new five-year plan that it would like to become a global supplier of these displays.³ This inevitably opens the door for new supply chain attacks and disruptions.

VR headsets will introduce a host of new vulnerabilities to an IT environment that will be a massive headache for the chief information security officer of any organization, let alone an individual consumer. The Augmented Reality for Enterprise Alliance (AREA) sums it up best by saying “Augmented Reality headsets open up new, unique, and significant threat potential to enterprise assets. They represent doorways through which bad actors can surveil, infiltrate, and potentially commandeer and misdirect critical resources and function.”⁴ Their reasoning for this is that no one wants to own the security problem for the headsets. As they put it: “There is a tendency for stakeholders to ‘pass the buck’ when it comes to taking responsibility for AR security: device vendors say it is the responsibility of the customer and can probably be handled by (mobile device management) MDM applications; MDM providers have not seen enough deployments to extend their platforms to meet AR-specific needs, which would not be sufficient in any case; AR project teams look to Enterprise IT for guidance; Enterprise IT and Mobility departments hesitate to open up their networks to these devices [and then do not develop guidance].”⁵

The lack of ownership is troublesome given the discovery of vulnerabilities in these headsets. A team at Rutgers University found that hackers could infiltrate a headset and “record subtle, speech-associated facial dynamics to steal sensitive information communicated via voice command, including credit card data and passwords.”⁶ Their research focused on vibrations captured by AR/VR headsets, such as “speech-associated facial movements, bone-borne vibrations, and airborne vibrations.” While seemingly innocuous, these vibrations could reveal detailed gender, identity, and speech information. These vulnerabilities could also allow bad actors to commit eavesdropping attacks, thereby enabling them to “derive simple speech content, including digits and words, to infer sensitive information, such as credit card numbers, Social Security Numbers, phone numbers, PIN numbers, transactions, birth dates, and passwords.”⁷

Security practitioners must also account for where and how these headsets are used. Compromised headsets used for video games present different risk levels than headsets used for educational, medical, or military purposes. Yet, the individual parts comprising the headsets may be ubiquitous across the headset ecosystem—regardless of the final assembler of the headset—and the discovery of a vulnerability in any given headset could have ramifications across multiple sectors.

Malicious actors could also exploit XR software to achieve their goals. A research team at Louisiana State University (LSU) found that bad actors could compromise a popular social and entertainment XR app and “take over a user’s VR headset, look at their screen, turn on their microphone, and install a virus on their computer.”⁸ The vulnerability also allowed any hacker to join private rooms, download additional malware on a user’s system, phish additional users, and send messages from a user’s account.⁹ Further, users interacting with the “infected” individual also caught the computer virus, facilitating a virtual pandemic.¹⁰ It is easy to imagine a scenario in which a bad actor extorts a victim by recording private sessions and interactions and threatens to release information unless ransoms are paid.¹¹ Even the most sophisticated and well-resourced companies are not immune from these potential cyber incidents. Security researchers found multiple vulnerabilities in Roblox that, if exploited, could unveil names and email addresses to millions of users.¹²

Regardless of the software manufacturer, malicious actors will likely find a way into the system and perform various types of cyber attacks. This includes launching an input attack that would prevent authorized commands from being recognized, making it impossible for users to operate in the metaverse. An adversary can launch a distributed denial of service attack and overload the system, resulting in network failures. Similarly, bad actors can perform a feedback overload attack that causes delays in information transmissions. An attacker could also overlay images into an environment “making it seem like new elements have been added or removed” or block the view in its entirety.¹³ The consequences of delayed or falsely overlaid information can be fatal during, for example, a medical operation that relies on XR capabilities.¹⁴

The authentication methods used to access a metaverse could also be corrupted or stolen, allowing unauthorized access or denying users’ access to their environment. But not all authentication compromises are created equal. A gamer’s inability to access Fortnite may not warrant a U.S. government response (much to the chagrin of the gaming community), but a soldier’s inability to train in STE may be of significant concern. Some XR devices, similar to smartphones and headsets, can only be accessed through certain biometrics, introducing another vector to capture this important piece of identity information. As airport security increasingly turns to facial recognition for screenings, for example, the integrity of someone’s facial biometrics is of vital importance.

Moreover, bad actors do not need to exploit a headset or the XR software to successfully impact an XR application. They could penetrate a data-hosting provider and launch a traditional ransomware attack that encrypts data necessary for an XR application to function.¹⁵ As a result, end users, like hospitals or the military, may lose the essential operational functions of their XR applications. In addition to ransomware, bad actors can steal user data from a cloud provider, corrupt data so that the XR application shows incorrect information, delete data, and provide a launch point to enter the XR environment or the user’s larger IT ecosystem.

While a hacker could use an existing attack method to inflict pain on a victim, they could also exploit new vulnerabilities within XR technology. This could include corrupting a digital twin, stealing virtual currency, or physically manipulating a user. A user’s avatar, or digital twin, could become corrupted or altered to an extent that makes it unusable.¹⁶ While defacing someone’s avatar may not seem like a national security threat, the mental and emotional toll on the user should not be overlooked. At a larger scale, if a city, such as Seoul, were to have its digital twin corrupted it may actually limit government services for its citizens. Although government services could still be requested in-person, the intent should be recognized for what it is: inflicting real-world consequences by debilitating the government's ability to provide basic functions to its citizens. Taking this a step further, a power plant or water utility could create a digital twin of their plant with schematics to help with equipment and sensor monitoring. A malicious actor could gain unauthorized access to that digital twin and could preplan physical or cyber attacks based on those digital schematics.¹⁷

One of the more unique outcomes that a bad actor could achieve is physically manipulating the user. Unlike most interactions with internet-enabled devices, XR technology’s primary purpose is to alter the reality of the user and to allow the user to interact in the metaverse by physically moving their bodies. As a result, a bad actor may seek to physically harm the user through a cyber attack. This has ominously been dubbed the “human joystick” phenomenon. In an IEEE study, they found that “87.5 percent of subjects were able to have their movements controlled by the addition of supplementary content to their VR screen…malicious software might add additional ‘objectives’ or mechanics to games that induce the user to move in a particular direction.”¹⁸

Unfortunately, these vulnerabilities already exist. In the same study where the LSU research team hacked a headset through an XR application, researchers also found that they could “disorient users and delete physical boundaries to make them walk into walls or fall down staircases in reality.”¹⁹ Another research team analyzed metaverse applications, systems, and traffic flow and found that an attacker could “easily mimic and visualize the immersive environment of the victim.” This could then lead to “integrity violations, content tampering, and the obstruction of users’ views” and cause “cybersickness.”²⁰ This phenomenon has already occurred in the real world when Pokémon GO reached its peak and criminals began abusing the game’s functionalities to prod gamers to walk into certain areas where they would get robbed of their belongings.²¹

All of the aforementioned vulnerabilities and exploits touch on various aspects of users’ and organizations’ privacy. In fact, there are several thoroughly detailed reports that focus on all aspects of privacy, and one literature review of “metaverse security” reports found that “most articles have examined the security of Metaverse environments from a privacy-preservation perspective.”²² These concerns are not misplaced, as XR applications will produce sensitive data about users that have previously not been generated—and they are ripe for the taking. The IEEE has published several reports on the matter and highlighted four key data captures that XR enables malicious actors to access that include movements and physical actions, neural activity, location tracking, and physiology.²³

This personal data does not include other information associated with a user or organization and is in addition to existing data tracked by basic internet use such as login credentials, personal identifiable information, financial information, or health information.

XR systems, like other technologies preceding it, will be targets of criminals and nation-state cyber attacks. The integration of XR systems and the data it produces becomes another opportunity on a widening attack surface. A criminal will deploy ransomware on a managed service provider (MSP) if they know it hosts the data for hospitals that cannot function without it. A nation-state actor may try to pilfer data from a server hosting DoD information that is generated from a fighter pilot’s XR headset to learn about an exercise that occurred in the South China Sea. Other malicious actors could preposition malware on an XR application that is widely used for the functioning of critical infrastructure operations in the United States and deploy it if hostilities were to increase. While the likelihood of these scenarios occurring varies, and their impact is dependent upon the integration of the XR device or application in a business unit, the U.S. government and private sector must be mindful of these legitimate cases.