Metaverse democratisation as a potential EA cause area
Epistemic status: half-baked at best.
Summary
In October 2021, Mark Zuckerberg announced the creation of a software infrastructure for a virtual universe that can closely interact with the physical world called the metaverse. The metaverse offers tremendous opportunities for social life, education and work. It also offers opportunities for large-scale manipulation and engineering addictive dependence. This post argues that we may now be at a tipping point that decides whether we are steering either towards a utopian or to a dystopian future of hybrid virtual/physical realities. It encourages a discussion on the assessment of the problem and brainstorming of potential solutions.
Background
In October 2021 Facebook was renamed to Meta. Their CEO Mark Zuckerberg announced to create the software infrastructure for a virtual universe, called the metaverse. As described in The Conversation, the “metaverse is an immersive virtual reality version of the internet where people can interact with digital objects and digital representations of themselves and others, and can move more or less freely from one virtual environment to another. It can also involve augmented reality, a blending of virtual and physical realities, both by representing people and objects from the physical world in the virtual and conversely by bringing the virtual into people’s perceptions of physical spaces. By donning virtual reality headsets or augmented reality glasses, people will be able to socialise, worship and work in environments where the boundaries between environments and between the digital and physical are permeable. In the metaverse, people will be able to find meaning and have experiences in concert with their offline lives.” Even if a significant fraction of apps and environments may not be created by Meta, they will likely build or depend on their infrastructure in one way or another. In other words, within a decade, we may have created a virtual universe that influences the physical universe. Unlike the fairly democratised structure of the internet, the power over the metaverse will largely be in the hands of a company with a strong track record on influencing elections and drawing people into addiction.
Opportunities
The metaverse could become a utopia where we
Stay in touch with friends and family abroad (or potentially AI avatars of deceased loved ones).
Do sports, go to concerts, museums, go shopping.
“Travel” for sightseeing, conferences, etc with close to zero carbon footprint.
Do practical courses on chemistry lab work, fire rescue, aircraft maintenance. Some of these courses can be cheaper to conduct, and some would allow practising of scenarios that are currently considered as too dangerous in physical environments.
Study science, as a more effective replacement to textbooks, thereby accelerating learning and ultimately scientific discovery.
Risks
The metaverse could also become a dystopia, where we
Do not have ownership of our virtual personality.
Have to pay money or data for virtual property that could be created for free in a democratised system.
Face environments that are engineered to be addictive.
Are subject to highly effective and scalable manipulation.
Are caught in a bubble that prevents exposure to a diverse set of opinions.
Most of the above points are true for present social media, but a metaverse will be social media on steroids.
Intervention points
I believe competing with Meta on a technological level will unlikely be successful due to narrower budget and potentially ethical constraints. However, one could look at cases where the internet adopted open protocols against the wishes of large companies and try to replicate this with metaverse protocols.
Beyond that, I consider governmental legislations as sensitive interventions points to guide metaverse developments into a utopian instead of dystopian future. These regulations may include:
Avatars must be based on open-source infrastructure. Everybody should be able to modify and disseminate the code avatars are based on, and everybody should be able to store all information associated with their avatar on their own hardware. Avatars that adhere to very basic single sign-on standards must be allowed into all public environments.
There must be substantial open-source infrastructure to create public environments and private environments for personal use.
There must be restrictions on what information can be shared with virtual environments.
There must be restrictions on what information about avatars/consumers can be shared to third party environments/entities.
There must be restrictions on actions that would be considered illegal in a physical environment. There is often no need for punishment, as such actions can often be technologically prevented.
Two barriers towards governmental legislation include (1) lack of technological sophistication in most governments to create and enforce large legislation, and (2) conflicts of interest in using the metaverse to influence elections. Both problems grow with the technological capabilities of a metaverse, making it critical to act early on.
Qualitative SNT statement
The scale of the problem of a de facto centrally powered metaverse would almost certainly exceed the importance of several election outcomes plus the damage done by current social media addiction. On the positive side, I struggle to estimate the counterfactual benefit of a democratised metaverse over its centrally governed counterpart.
Given current attention to the COVID19 pandemic and several environmental crises, I have not yet seen any parliamentary discussions on metaverse regulations. Democratising the metaverse currently seems highly neglected.
Laws like the EU General Data Protection Regulation and the California Consumer Privacy Act demonstrate the tractability of implementing regulations. I am uncertain about their effectiveness compared to zero regulations and a hypothetical loophole-free system.
Call to action
Please comment below on whether you think democratisation of the metaverse is a worthwhile cause area
If you think so, please help me to improve this post by suggesting/commenting on its Google Docs counterpart.
If you and the community think metaverse democratisation is a worthwhile cause area, please share information on this issue with friends in relevant positions.
Acknowledgements: Thanks to EA Graz for the discussion on the topic and Aditya Vaze for his feedback on the draft.
I think the risk of a VR monopoly is probably low. A metaverse is just an office-compatible VRChat with external app screensharing, and portals. Lots of people are capable of making those. It’s also not apparent to me that Meta are dangerously better at making the hardware than anyone else (apple, HTC and varjo are all stern competitors. Even HP are in the game. Also, by the way, you personally would probably be interested in SimulaVR, they’re about to start making the first portable, self-contained VR workstation computer, and it runs NixOS. It is going to be so so much cooler and better than a laptop. I don’t think I’ll be buying it, it’s only 30 pixels per degree, the maximum perceptible, 60, is right around the corner. But I really hope they stay in business).
I don’t see any routes to lock people into only using a single service for VR socialization either. You’re concerned about exclusivity with user avatars, having to pay for virtual objects or access to locations, but all of those things make the platform shittier and less useful, if they do that they’ll get outcompeted, there’s nothing they can currently do/are doing to prevent others from competing with them in that way.
There’s just way too much competition and incentive for collaboration on standards here. It was heartening to me to read Unreal’s arguments on this. I’d guess that game engines have a big role to play here, and the engines are getting pretty girthy at this point, substituting Unreal out for something made in house is not really going to be feasible for Meta, but Unreal has no interest in aiding the formation of a hardware platform monopoly, so Meta is going to be strongly encouraged to support Unreal’s open standards work.
I think your political thinking is pretty vague. Democratization and addiction/misuse/inhumane tech also have very little to do with each other, if anything they’re aligned. The more democratized the system is, the more addictive systems pop up and compete within it, and the faster their victims find their way to them, one barrier removed.
Platform monopolies don’t have a greater incentive to promote addiction than an open market already has, and in practice, platforms are usually not the drivers of addiction within their walls. Twitter is a good counter-example, it emerged on the open web, it consisted of some ridiculously crude, barely designed mechanics that worked for reasons nobody understood, and then Twitter went on to become the dominant platform for global discourse, causing a huge amount of damage to the world. The web platform didn’t do that. The openness of the platform didn’t prevent it. Twitter wasn’t addictive because it was carefully designed to be, it just was addictive. And twitter isn’t a monopolistic platform either! It has a simple business model and no lock-in. Your data on twitter is all public. Anyone could take it and make another twitter. You can take it and leave. Sometimes people do leave, but many of them go to mastodon, which is exactly the same, and has the same discourse-political problems.
The inhumane pattern of twitter persists because addiction is driven by voluntary choices. You can’t solve that sort of problem by adding more liberalization.
You could only improve the situation by creating platforms that are more engaging, but don’t cause the same harms.
You need a constructive vision! Give people something better!
The root cause of the problems are badly designed systems, or an absence of any good designs. People don’t want to design bad systems. They just do whatever pops into their heads.
I don’t particularly like facebook, but I don’t think they want to destroy democracy or create an unlivable world. No one has those values. I think we need to take care not to get too cynical. Cynicism is a terrible vice. It doesn’t just help you to identify your enemies, it creates enmity and it keeps the rift from ever healing.
But given the opportunity to lobby Meta, I wouldn’t know what to say to them, yet. Systems determine outcomes but no one is offering a theory of VR hangout and networking systems in relation to how people get information, or have political conversations, or find work, or find love. There are lots of great challenges in designing humane online systems. It’s not obvious that any of the big steps we need to take have much to do with VR. Probably some of them do, but I don’t know what that looks like.
I am responsible for some visions in social computing (linked: A technology for larger yet more focused communities). Even though they can’t obviously be situated in VR, I’ll mention them just in case they spark any ideas with others.
Most of my hopes are with venues like Roamresearch, places where it’s as frictionless and accessible as possible to edit linked, structured information collaboratively. Instead of promoting ephemeral, viral content, it promotes longer, continuously evolving content maintained by passionate, curious research and curation communities.
I think these things could easily evolve into general purpose social network infrastructures too.
There’s a parallel question there, should we promoting, or, hell, funding, Athensresearch, the open source alternative to Roam? It’s possible that we should! Roam currently doesn’t seem to have plans to support forum and chat primitives, while with Athens it seems like those things can already be added, by anyone, it could end up catering to a much broader, more globally impactful usecase. However, there are some big technical challenges that need to be solved for supporting globally connected versions of these systems, and I don’t really see either project engaging with them yet.
I wouldn’t call it ‘the’ open source alternative to Roam, since the wording suggests it is the only such alternative. I personally use and recommend org-roam.
org-roam doesn’t appear to be multi user
Wonderful comment.
Nice
Thanks for sharing your insights Mako! After reading your response and the IEEE Spectrum article you mentioned, I am much more optimistic that the metaverse can/will move in the right direction. Is there anything that could be done (by governments, companies, NGOs, the general public, or whatever player) to make this even more likely?
I also liked your example of Twitter, where addictiveness was not designed into the system, but happened accidentally. Accidents usually prompt investigations to improve regulations, for instance in the aircraft industry. Do you think there are any concrete key learnings from the case Twitter how to prevent similar accidents in the future of the internet or metaverse? If so, could or should some of these be baked into better designs, and are current incentives aligned with this or would it require some governmental regulations (since you are worried about liberalisation)?
I still believe that Meta is a major player on the market. And while I do agree that they have no direct interest in destroying democracy or creating an unliveable world, I think they act in line of Milton Friedman and would do just try to maximise their profits. I am not sure if there is anything wrong with that in principle, as long as the rules of the game ensure that maximizing profits aligns well with overall utility. In the past, I don’t think the rules of the social media game aligned well with overall utility. And I am not sure that the need for and support of open standard by players like Meta alone is sufficient to align profit maximization with overall utility in the metaverse. If this assessment is correct, it would make sense to brainstorm ideas for such an alignment as the metaverse develops.
Btw. thanks also for sharing your LW article on Webs of Trust (on my reading list) and your thoughts on RoamResearch (pm’d you with a question on Roam vs. Obsidian).
Fair prompt. I get the impression that the most impactful thing you can do is to make sure that the people leading the standards dialog have strong technical vision and good taste. That’ll also make it more likely to even succeed at establishing a standard. I guess that’s something that EA (with so much software engineering acumen) could probably do better than most NGOs! But yeah, it looks like that might already be the case, I’m not sure.
I don’t know what the addictive social media systems of VR will look like. It might just be twitter again, but with bigger text.
Hmm… I guess VR social systems might orient around VR’s adaptation to voice chats, ubiquity of mics, support for body language (filtered through an avatar, which will often make people more comfortable) and a more natural sense of presence.
I find it difficult to imagine many novel systems about that, because it seems like it’s constrained to the sorts of arrangements that’re already pretty natural for humans. People walking around in a room and making sounds at each other. If you’re rude, people remember, and you don’t get invited next time. It doesn’t seem obvious that the information or the social bonds can be structured in any alarmingly novel ways. Well, I guess one big difference is that the social cliques can end up a lot more globe-sprawling and specific and extreme. But I’m not sure. There will still be lots of cross-linking. You’ll tend to meet your friends’ friends.
Maybe systems will end up being.. less about structuring information, and more about structuring relationships, controlling group matchmaking or timetabling.
I don’t think this is a cause area but do think this is a sub cause area within longtermism. I think if we do want to be able to upload our consciousnesses and live as digital people, the initial formation of the metaverse could pose lock-in effects that could make life not great for future digital people (so I guess more of an s-risk concern). I’m actually already working on a project for this and am glad to see other EAs thinking similarly. Sadly I cannot comment more on it though since I work for the government.
Good to hear that there are EAs working on that within governments.
This strikes me as being at least as plausible a cause area as the idea of working on legal frameworks governing space colonization, insofar that we think near-future space exploration or metaverse efforts will have some influence on the shape of far-future galaxy colonization or virtual life in an “age of em” style scenario.
Perhaps we should advocate for the adoption of open-source metaverse infrastructure. This approach allows users to modify and disseminate the code underlying avatars, public environments, and protocols. It ensures that the metaverse remains accessible and customizable, encouraging active community participation.
Problems around social media (which are similar as you say) are half-mainstream in the media today, such as in Black Mirror.
I agree this seems important, but also It’s a topic that people are aware of.
I would personally be happy to see an EA take on it (probably worth one person researching it at least to come up with general plans or listing existing ones), though it seems to require expertise which goes beyond my own.
Anyway, you have my upvote
My own take:
Defining a good regulation for this kind of problem is complicated, even if we could arbitrarily write laws (but assuming we could do so only once).
For example, it seems to me like “you must allow avatars into your environment” would be wrong:
Big companies could DDOS small competitors with avatars that the small competitors will have to accept
I think that defining the data as public is better. Imagine if any company could create their own UI for Facebook, and have all the people and posts available.
I’m aware that defining such a data structure is not trivial, but it seems like a better direction to me
There are open source social networks that have a similar vision to this in mind. Part of the hard problem of starting a new social network is coordinating everyone to move to it. Perhaps EAs would have something smart to suggest around there
If I am at least partially right about this, then it isn’t a huge $100M project. The hard parts might be solved by a few smart people thinking about it together, and then they’ll publish the solution for existing open source social networks to adopt. I don’t know. That is why I’d personally have one person flesh out directions as a first step, and I wouldn’t reject this project just because more important (1000x more expensive) projects exist (which will often need different people working on them. This here isn’t a math problem, I think)