Metaphysic.ai
Published in

Metaphysic.ai

Ethics in the Hyperreal Metaverse

The metaverse will become hyperreal. Today, AI tools can create synthetic content that is indistinguishable from video shot with a camera. Soon, video games will be so realistic that they look like live-action movies and Zoom calls will be like the Matrix where photo-realistic synthetic versions of ourselves interact inside 3D immersive worlds. When our experience of the metaverse feels real, it will become an extension of reality. This article sets out the issues I think about when designing products for a hyperreal metaverse.

Our online and offline worlds are converging towards a single experience where highly personalized, immersive digital content will become a seamless extension of real life. As this hyperreal metaverse emerges, it is imperative that we embrace strong ethical principles that will guide its development in a direction that is fair, inclusive, and representative of the human experience in all its diversity. As our virtual and physical identities entangle, we must build systems that empower individuals to control the fundamentals of who they become in the metaverse. These systems must complement and extend our highest ideals of equality, individual sovereignty, and liberal democratic institutions in the real world.

The confluence of several technologies — AI, blockchain, and XR (mixed reality) — has laid the foundation for a hyperreal metaverse that holds immense promise to transform the ways we work, play, and connect with others. But the metaverse is also rife with technological and social challenges that must be overcome to unlock its full potential. The evolution of the physical/digital divide towards reality/hyperreality will mark a profound transformation in human history on par with the industrial revolution. Just as we gave up the rigor of agrarian life for the chaos and capital efficiency of the big city, it is inevitable that we will increasingly spend time fully immersed in photo-realistic virtual worlds that feel more and more like home.

I believe that the people involved in building the metaverse have a responsibility to think critically about the technology’s impact on who we are and how we interact with society. My co-founders and the team at Metaphysic share this perspective. We are leaders in using AI to make deeply personal content that looks real. This format of content creation can be scaled to every person that uses the internet and soon we will be able to build hyperreal worlds based entirely on an individual’s unique perception of reality — and the visual data it entails. We founded Metaphysic on a commitment to ethical leadership and we understand the need for strong community involvement in developing the technology. To drive public awareness of these issues, we created Synthetic Futures, an organization that brings together industry leaders, policymakers, technologists, and content creators in an open forum to discuss the opportunities and challenges that await us in the metaverse.

We launched Every Anyone to empower individuals to control their own hyperreal identities and safeguard their biometric data. At the core of Every Anyone are blockchain and Web3 infrastructure that leverages AI to create hyperreal content directly from data owned by members of the community. Every Anyone’s mission is to reconfigure how regular people participate in the internet and the economies that drive its development. This is an ethical and moral necessity as the hyperreal metaverse expands to reflect the real world with greater and greater granularity. It is also a testament to the power of web3 systems to put individual users at the heart of the online experience.

Technical innovation aside, it is the choices we make as developers, users, policymakers, citizens, neighbors, and individuals that will have the biggest impact on whether the metaverse is a net positive for humanity. It’s clear that robust ethical frameworks are a prerequisite for making better choices in building the metaverse that we want to inhabit. But I am the first to admit that we can’t predict how this space will evolve. In this sense, ethical leadership is not only about creating codes of conduct to guide our actions — it is about fostering an environment where we can all engage in continuous evaluation and values-based discussion about our vision of the future.

Discussion Topics for an Ethical Hyperreal Metaverse

I have set out some of the themes we discuss when building products for the hyperreal metaverse and collaborating with creators to make photorealistic deepfake content. It is not an exhaustive list. My aim is to provoke discussion among those building AI content creation tools and the wider base of stakeholders who will be inhabiting the metaverse we build together.

Deep Immersion and New Types of Harm

Synthetic content in a hyperreal metaverse will look real and participants may believe, if only for a moment, that the content created by AI was actually shot in physical reality with a camera. This content might include a hyperreal version of yourself that is generated by AI using biometric data such as photos of your face and body or recordings of your voice. This content will be created and hosted by metaverse platforms, gaming companies, content production studios, and other service providers. There is clearly a potential for harm at the individual and social level if these third parties can make hyperrealistic content of you doing or saying something without your direct physical input. These potential challenges need to be addressed now — not once they’ve become a problem.

Hyperreal content consumed in the metaverse via mixed-reality devices will be deeply immersive. The experience will feel ‘real’ on a deeper level than navigating a game-like fictional environment while wearing a VR headset. We need to study the impact that this format of sensory immersion has on our mental health and sense of self. Some examples include:

  • Research has shown that using VR content for people with post-traumatic stress disorder (PTSD) can have profoundly positive results. It is not difficult to imagine the counterfactual, where abuse experienced in a hyperreal, multi-sensory environment may be more traumatic than it would be in less realistic digital contexts.
  • We do not fully understand the impact of immersive experiences on the developing minds of young adults. VR content has demonstrated that it can be a powerful tool for improving educational and social experiences. But these technologies are so new that it is unclear how they may influence the cognitive and emotional development of young people in the long run.
  • Another issue is the degree to which XR devices and headsets will need to collect large amounts of personal data about our faces, emotional responses, and private spaces like our homes. It is critical that we create robust consensual mechanisms that allow the individual to determine the degree of access that governments and companies have to our state of mind and physical lives.

It is clear that we need to proceed deliberately and with caution when developing and bringing hyperreal content and associated technologies to market. There are immense benefits to a hyperreal metaverse in terms of connecting people through shared experiences, especially when it comes to increasing empathy by creating a virtual environment that enables authentic human emotion. I believe that these are important outcomes for healing some of the social divides, political partisanship, and ideological radicalization we have seen play out on Web2 and social media platforms. If a deeply personal hyperreal metaverse is going to be a positive step in the evolution of the internet, its careful and ethical development is paramount.

Individual Agency

As a society, we extol the privilege of exercising unencumbered agency over our physical, verbal, and mental capacities. Unfortunately, most people around the world experience some form of constraint over their individual agency, whether it's coercive economic systems, autocrats, bullies, abusers, retrograde institutions, physical walls, or discrimination based on gender, race, or sexual orientation. The situation is further complicated for people with neurodivergence or those diagnosed with mental illness, which may be unjustly used as grounds to constrain their autonomy. There is no end to human cruelty or the challenges created by our own bodies. Everyone experiences these challenges differently, but I believe that the majority of people hope for more control over who they are as individuals. We launched Every Anyone because a globally connected hyperreal metaverse can put individuals at the center of a new internet where they can control how their biometric data is used to create personalized content and shape their identity. We must design systems that expand individual agency and autonomy over our hyperreal identities in the metaverse so we can transcend the limitations we experience in the physical world.

The metaverse is both lauded and derided as an escapist utopia where individuals can shake off the constraints of the physical world and become anything they want. What does that mean? Often, it means the chance to have a cartoon avatar inside a virtual world that is designed, owned, and operated by a giant corporation for the benefit of anonymous shareholders.

These types of arrangements have made the Web2 experience incredibly convenient, but they have also obfuscated their costs and risks. As technology evolves to a point where third parties can recreate what makes us uniquely individual without our consent, we must demand more agency, participation, and control over ourselves in the metaverse as it develops.

Consent

Who we are in the metaverse will become who we are, period. We need to build systems that allow individual users to have informed consent over how their biometric data is used, where their hyperreal likeness appears, and how it acts. An important feature of individual agency is the ability to grant and withdraw consent. All of our social, political, and economic systems in physical reality are ultimately predicated on the assumption that we are in control of ourselves and consent to participation. In an environment where artificial intelligence can alter identities at scale, individual choice is paramount. The challenge is that the mechanics of algorithmic systems are often illegible. Consent depends on both the absence of coercion and an informed understanding of the terms. But AI systems can manipulate users in subtle ways and processes are often buried deep within the terms of service. A core feature of Every Anyone is to establish an ethical, transparent economy for content creation and participation in the metaverse where individual users can grant informed consent that creators can rely on when bringing personally identifiable data into their creations.

Southpark Studio’s deepfake satire show Sassy Justice mocks many well known celebrities and American politicians. Credit: Southpark Studios

Diversity

As it grows into a seamless extension of physical reality, the metaverse must be designed to reflect the limitless variety of human experience. This requires the development of datasets and models that are intrinsically inclusive and dynamic enough to handle fluid notions of personal identity. Collecting personally identifiable data for complex models ethically with full licensing is incredibly hard today. As a result, best in breed datasets are small and invariably introduce algorithmic bias that systematically misrepresents certain groups of individuals. As the metaverse becomes more important in our social and economic lives, limiting the potential for algorithmic bias is essential. Through thoughtful design of viable bottom-up data marketplaces and the intentional development of XR technologies that cater to users of all backgrounds, we can build a strong foundation for a just virtual world. Every Anyone is built on Web3 infrastructure that can facilitate micropayments and track the use of an individual’s data. We envision that our future virtual lives will be powered by datasets populated by people from our communities who also have an economic stake in the use of their data. Diversity is the key to any healthy ecosystem and the metaverse is no exception. We hope that our commitment to diversity and inclusion will, in some small way, contribute to a metaverse that is an open and welcoming environment for everyone.

Open/Closed Systems

The tension between closed and open systems will be important in how individuals’ rights and agency are treated in the hyperreal metaverse. Ultimately, I believe that the metaverse will be a constellation of open and closed platforms and services where corporations do their best to lock users into their ecosystems while Web3 services and protocols facilitate portability across experiences that will be increasingly popular as the offerings of smaller, independent developers mature. As the tech giants that dominate our internet experience today migrate their services to the metaverse, it is likely that they will invest heavily in content moderation and tools to mitigate user harm. On the other hand, the proliferation of smaller, independent content platforms may cater to market demand for more permissive virtual environments where content and user behavior are not moderated. Past experience with open virtual worlds has shown us that these systems can be exploited by bad actors who undermine the safety and enjoyment of others inhabiting the space. This can take a variety of forms including sexual abuse and hate toward users on the basis of their sexual orientation, gender identity, or race. In the case of hyperreal content, the detrimental effects of such actions are magnified as a result of the close link between a user’s physical and virtual identities. The development of a closed metaverse where the actions of users are intentionally constrained through technological decisions can reduce the negative effects of bad actors, but this comes at the cost of limiting freedom of expression and artificially limiting its creative potential. Technologists, content creators, and users must have an open and frank discussion of these tradeoffs to find a balance between freedom and safety that privileges the fundamental rights we cherish in the physical world.

The documentary Welcome to Chechnya used synthetic faceswapping to protect the identity of persecuted gay Chechnyans while preserving their humanity. Credit: Welcome To Chechnya

Moderation and Gatekeepers

On Web2 platforms, like social networks, many of the potential harms discussed above are addressed through centralized content moderation. The challenge is that platform operators become gatekeepers with definitive authority over user behavior — from the time any individual person logs in they become ‘users’ in a system totally governed by the platform operator. This authority is necessarily directed to optimize shareholder returns, sometimes to the detriment of the fundamental rights of individual users. When it comes to the hyperreal metaverse, we need to discuss the relationship between individuals and organizations — be they corporations or DAOs — that control who creates content and how it is distributed. The stakes are high in a hyperreal metaverse that relies on the biometric and other privacy-sensitive data of individual users that is imported from the real world. It is essential that we deeply consider how content moderation policies and other regulatory frameworks may impact users in the hyperreal metaverse and take proactive steps to limit these potential harms.

Embedding Negative Social Structures in a Metaverse for Everyone

The hyperreal metaverse will likely be an extension of our physical world that allows individual users to enhance, edit, or invent new features and layer them on top of an imported hyperreal version of the real world. While it is technically and logistically easier to scale a hyperreal, personalized metaverse to billions of users by effectively making a contextual copy of the real world, there are negative power structures and institutions that we must work hard not to replicate. We must make conscious design decisions that attempt to avoid proliferating systems that embed social and economic inequality. For example, when designing Every Anyone’s Community Collection of 10k fictional people, we found that certain minority groups were not adequately represented in our training dataset of approx. 100k unique faces of real people. As a result, we had to create over 50k fictional people and edit down to a selection of 10k that is a closer representation of people in the world today. The collection still does not perfectly reflect all the different groups of people that exist in the world, especially minorities. Through Every Anyone, we hope that we can engage people from everywhere on earth and license their data ethically so that future AI models are adequately representative.

There is also the risk that new class systems will emerge in the metaverse where the haves and have-nots are stratified by their avatar’s access to virtual fashion items and other virtual modifiers. There is no doubt that the most desirable NFTs today are expensive and totally inaccessible to regular people. Our goal for Every Anyone is to provide a level playing field where anyone can mint their own hyperreal avatar for free (plus minting costs). But we need to work together with other organizations building the metaverse, including NFT creators and virtual fashion houses just to name a few, to give regular people access to desirable virtual appearances, accessories, and experiences.

Labelling Manipulated and Hyperreal Content

In the future, hyperreal content in the metaverse will be indistinguishable from footage shot with a camera and XR technologies and wearables will immerse viewers to an extent that content feels real. While this creates exciting new opportunities for creators to delight their audiences, there is also the potential that it will be used by bad actors to deceive, mislead, or defraud. This raises important questions about the need for disclosure and labelling of both hyperreal and manipulated content and under what circumstances disclosure is warranted. A good example of this tension is hyperreal representations of politicians and political messages in the metaverse. An informed electorate needs high-quality and accurate information for democracy to function as intended. Does this mean that politicians should be required to faithfully reflect their real-world identities and actions in the metaverse and disclose when they deviate from this baseline? If so, who gets to decide and how will these policies be enforced? These questions don’t have easy answers, but it is important that they are at the forefront of our considerations in building an internet powered by hyperreal content.

Viral sensation DeepTomCruise used synthetic de-aging of the Hollywood start and labelling to make it clear the content wasn’t real, but viewers still occasionally believe it is the real actor.

Building an Ethical Foundation

The ethical tensions set out above are not intended to be an exhaustive or prescriptive list. Creating an ethical framework for the hyperreal metaverse is not an exercise in labeling some things as good and others as evil. Rather, it is an open discussion about the pillars of a new social contract that will shape our actions in virtual public spaces. Just as the metaverse will open new creative possibilities that we can’t yet imagine, so too will it create new ethical quandaries that our current frameworks aren’t equipped to handle. While we can’t anticipate all the ways that the metaverse will evolve, it is critical that we begin to explore its ethical implications now so that we are prepared when new issues arise. With these ideas in mind, we have designed Every Anyone to explore issues of personal identity in the metaverse and we are committed to open dialogue in the future. By joining the community at Every Anyone, you become a stakeholder in a global experiment that is trying to understand what it means to be human in an increasingly virtual reality. Only by working together can we build a hyperreal metaverse that embodies our highest social ideals and the best elements of human creativity while elevating our collective potential to new heights.

Metaphysic builds software to help creators make incredible content with the help of artificial intelligence. Find out more: www.metaphysic.ai

For more info please contact info@metaphysic.ai or press@metaphysic.ai

About the author: in a galaxy far away, I was a lawyer turned internet & society researcher. In the 7 years before co-founding Metaphysic, I built tech companies in SF and London. I have always been obsessed with computational photography and computer vision, so it is a thrill to work alongside amazing people on the next evolution in how we build and perceive reality — one pixel at a time.

© Thomas Graham & Metaphysic Limited 2021.

--

--

Get the Medium app

A button that says 'Download on the App Store', and if clicked it will lead you to the iOS App store
A button that says 'Get it on, Google Play', and if clicked it will lead you to the Google Play store