Fullstack Killswitch

Fullstack Killswitch. Towards a Integrated ‘Universal Escape Key’ for All Extended Reality Systems
Mathana Stender, May 2020
In the 1975 classic ‘Bohemian Rhapsody’, Freddie Mercury asked “Is this the real life? Is this just fantasy?” With the rise of immersive, ‘extended reality’ (XR) technologies like augmented reality (AR), virtual reality (VR) and brain-computer interfaces (BCI), this question of real-life-versus-fantasy takes on a new imperative. XR technologies will soon impact many fields and disciplines such as macro economics and labor, brain science, research ethics and education systems design.
The novel approach presented here is that of a “universal escape key” (UEK) that serves as a trigger function to allow a user to instantaneously escape from an immersive environment via a pre-designated input control that propagates across the entire fullstack framework. Like Windows’ Ctrl+Alt+Delete or MacOS’ Command+Option+Esc hotkeys, an XR UEK would promptly end an XR experience. Such a UEK would need to be engineered into the fabric of all new XR-supporting systems, subsystems and compatible ancillary systems.
Responsible Research: Charting The Unknowns
The still-unknown effects of long-term exposure to extended reality environments on the human brain (eg. neurological ontology and cognitive epistemology) and XR-specific health side-effects like “simulator sickness”1 and acute, momentary paralysis give rise to the need for a standardized ‘escape button’ functionality across XR user control interfaces regardless of manufacturer, platform or environment.
Most of today’s large online social platforms seek to “self-regulate”, meaning giving themselves the power to write the rules of their terms of service, enforce content and take action against accounts that they find to violate the rules2. Content reviewers have developed serious psychological trauma3, and advertisers have been banned from “shocking” ads4 to protect users. But as XR technologies progress, and the ecosystem fractures from the hegonomic hold of the the major technology companies, the new sensory experiences that we are exposed to become potentially more severe to our cognitive development and psychological wellbeing.
Many questions remain in regards to the interplay between evolutionary biology (nature) and cognitive flexibility (nurture), and while this interplay between our minds and the world around us is still a hotly debated topic in psychology5 and philosophy, developments in extended reality computing add additional layers of obfuscation between the “real” and the “perceived”. As we are witnessing the dawn of a new type of technologically-mediated experience, that are profound ethical, a philosophical and physiological question that arises: can a new type of immersive experience that simulates a new kind of “reality” alter the foundations of one’s epistemic and ontological perspective of the very nature of truth, perception of self and understanding of the other.
Health, Wellbeing and Safety Considerations

  • Stimulation-induced temporary paralysis from visual stimuli could lead to the momentary loss of some motor functions. Certain groups, such as those with epilepsy or other neurological conditions are at higher risk.
  • Those with pre-existing physical conditions may not be able to simply “take off” a headset. Furthermore, as XR systems become more invasive (like theoretical neuro-assisted computing6) there may one day be no external headset to remove.
  • A universal escape key therefore creates a hardcoded safety mechanism that preempts a generation of XR technology from having to retrofit solutions for non-able-bodied users. To avoid ableist designs creating a barrier to user experience, and create XR systems predicated upon ethically-aligned design principles, a global kill switch in the form of a user-defined escape key is essential.

Break-Glass-Killswitch: A Universal Escape Key for Extended Reality Environments 
The internet was invented just over 50 years ago, and during the last half-century, exponential technological development has developed piecemeal, adding an increasingly interconnected network of information sharing services beyond the wildest dreams of its creators, both in global transformation but also destructive capacity7. The development of immersive technologies like AR, VR and brain-computer interfaces (BCIs) provides a chance to create a proactive vision for a new type of human augmented computation, one that is rooted in human-centric design principles. The ability to shut down an immersive, extended reality is an engineering feat that should built into XR frameworks from the beginning of their commercial applications, and not as a retrofitted patch to deal with ethical dilemmas once they have already revealed themselves. There is, therefore, a compelling and urgent safety interest for a “universal escape key” (UEK) to be incorporated into all technical systems that display individual optical or neurological immersion. This escape key functions like a master kill switch, and when triggered, would “crash” the entire immersive system, bringing the users back to a visual reality unmediated by immersive elements. On a practical technical level, the user would set up a predefined action (gesture, ‘safe word’, or other stimuli) before entering an XR environment (this could, for example, accompany the on-boarding phase where users set gyroscopic calibration of input controllers) that could be triggered at any point.
It would require standardization of hardware protocols (chip architecture), software parallelization (root level) and rendering engines (static, dynamic and algorithmic) that powers immersive extended reality (XR) environments. Technical standards, UI/UX calibration, ethical best-practices, safety-focused design of future systems and tools like a standardized, open APIs are key to streamline interoperability. But immersive reality systems are too new to fully understand their full potential or to what forms of interaction to which they’ll scale. A killswitch, that propagates throughout the entire immersive environment stack, provides a conscious design choice to put the user in control of whether or not, from one moment to the next, they want to be immersed.
Industry Prerogatives vs. Societal Imperatives
As the result of a tech R&D culture and market ecosystem in which proprietary builds are most common, there is a lack of standardized input control functions throughout XR ecosystems, so hurdles to this framework do exist:

  • Considering some XR operating systems are built upon legacy code libraries, meaning such a solution might require substantial retrofitting and reverse engineering. 
  • A lack of standardized input control functions means that user control is constrained to proprietary input. This approach should also account for idiosyncratic differences in UI controls for the same environment (e.g. a VR game) across multiple platforms (e.g., Playstation or Xbox).
  • The 3D physics rendering systems that power immersive game environments is somewhat of a duopoly8. Buy-in from both Unity Technologies and Unreal Engine is essential. As virtual rendering engines pivot from games to other forms of interactive experiences, the robust developer communities9 will become stakeholders in ensuring seamless adoption of the UEK.
  • Furthermore, some proprietary headsets are created by companies that control a significant portion of the content-delivery chain. The Oculus Rift, while a headset, is part of an integrated corporate ecosystem that acts as hardware manufacture, game platform and content hosts; whereas products like HTC Vive headset can be tethered to various consoles. Creating a seamless integration for the UEK may very well require new technical standards and protocols, regulatory oversight mechanisms and industry buy-in at the hardware, software and platform levels.
  • UX/UI interfaces and content delivery (i.e.,  computation run on-device versus in the cloud) throughout the AR/VR ecosystem create a hurdle for interoperability for a “global command” that is recognized by all devices, operating systems, platforms, and episodic environments (e.g., games, interactive films, etc).
  • Regulatory regimes will be needed to ensure compliance, and new laws created to create (dis)incentives for duty of care. As European data protection laws like the GDPR have shown, sometimes a single regulatory ‘zone’ can force the tech industry to create new user protections that they would not pursue on their own. A market with a large enough consumer base could therefore theoretically drive the development of an XR UEK via legislation that would impact user-facing system design for the entire world.10

Threat Modeling the Unknown
Malicious Design, Subverted Realities

  • “Entertainment” is a broad category of stimulation, and is in many ways very subjective. As suspense and horror genres of films and games take on a new viscerality, a user could become overwhelmed by fear during  the experience. While XR experience may one day have ‘ratings’ like films and video games, a board akin to the Motion Picture Association of America (MPAA; the U.S. film rating body) won’t necessarily take into account the ‘scare factor’ for entertainment products that are designed to frighten.
  • ‘Dark patterns’: App and website design has been used to obfuscate functionality and nudge users into clicks. Given that early creators of XR technologies that power immersive environments are many of the same hardware manufacturers, content platforms and “game” engines that currently dominate the tech landscape, we can extrapolate that the financial incentives and desire to consolidate market shares will be path dependent on current market key performance indicators (KPIs) and normative quantification of returns on investment (ROI). 
  • Current ROI incentives based on contemporary KPIs include engagement metrics and advertiser impressions, and while new forms of revenue-driven indicators will undoubtedly emerge, the current applications of XR technologies are likely to use design features similar to those used in 2D virtual environments that seek to keep users immersed in a virtual experience.

Retrofitted Preemption & Prevention 
Future brain-computer interfaces11 and human-to-human interfaces12 that contain neurological components are a theoretical case study in the need for a UEK. Being able to safely exit a brain computer interface will be particularly important. A UEK should be engineered to ensure that the crashing of one layer doesn’t have an adverse effect on another layer. If the neural stimuli and contact electrodes are not able to shut down in unison, a user could be left with adverse side effects. Health issues like simulator sickness, along with the unknown impact of long-term immersion in a VR environment13 create an impetus to create the technology with a human control as a guiding principle. While being able to abruptly exit an AR overlay may not seem like one of the most pressing ethical issues facing technological development, if scaled to neurological inputs and cognitive computing, the need for such a “break glass in case of emergency” protocol becomes much more apparent. 
The 2020 COVID19 pandemic will hasten the adoption of XR technologies across labor markets and critical social industries like education and healthcare. As more societal functions will be brought into the ‘virtual’, it is imperative that hardcoded escape functionality be instituted for all XR systems.  Technologists now have a chance to hardwire human-centered design into an entire class of technology. This may require creating engineering solutions, but if we catch it now, humanity can avoid the myriad problems and pitfalls that could besiege this nascent technology. 
Mathana is CIHR fellow and a Berlin-based tech ethicist who investigates the socio-cultural impact of AI/autonomous systems, AR/VR and other emerging technologies. They have recently co-authored the ‘Extended Reality’ chapter in Ethically Aligned Design, an ambitious best-practice industry guide published by the IEEE Global Initiative on Ethics of Autonomous and Intelligent Systems.


  1. Davis, Simon, Keith Nesbitt, and Eugene Nalivaiko. 2014. “Systematic Review of Cybersickness.” In ACM International Conference Proceeding Series, 02-03-December: 1–9. https://doi.org/10.1145/2677758.2677780
  2. York, Jillian C., 7 April 2020. The global impact of content moderation, Article 19.
  3.  Newton, Casey, 12 May, 2020, Facebook will pay $52 million in settlement with moderators who developed PTSD on the job. The Verge
  4. YouTube, 2020, Violent and Shocking Content in Ads. Google Support Center
  5. Srinivasan, Amia. 2015. Normativity Without Cartesian Privilege. Philosophical Issues. 25. 10.1111/phis.12059. 
  6.  Lopatto, Elizabeth. 2019. ‘Elon Musk unveils Neuralink’s plans for brain-reading ‘threads’ and a robot to insert them; The Verge
  7.  Kleinrock, L. Oct. 29, 2019. Opinion: 50 years ago, I helped invent the internet. How did it go so wrong? LA Times
  8. CBO Insights. Sept. 2018. The $120B Gaming Industry Is Being Built On The Backs Of These Two Engines.
  9. Dillet, Romain. Sept 5, 2018. Unity CEO says half of all games are built on Unity. Tech Crunch
  10.  Albrecht, Jan Philipp. “How the GDPR Will Change the World.” European Data Protection Law Review 2 (2016): 287-289.
  11. Degenhart, A.D., Bishop, W.E., Oby, E.R. et al. Stabilization of a brain–computer interface via the alignment of low-dimensional spaces of neural activity. Nat Biomed Eng (2020).
  12. Jiang, L., Stocco, A., Losey, D.M. et al. BrainNet: A Multi-Person Brain-to-Brain Interface for Direct Collaboration Between Brains. Sci Rep 9, 6115 (2019)
  13. Ahn, Sun Joo, Jeremy N. Bailenson and Dooyeon Park. “Short- and long-term effects of embodied experiences in immersive virtual environments on environmental locus of control and behavior.” Computers in Human Behavior 39 (2014): 235-245.