Professor, Designer, Husband, Father, Gamer, Bagpiper

Paul Hoover posted a Mirrorworld Bill of Rights recently, and a group of people got together in Hubs to discuss ideas related to AR and the kinds of protections people should have when it becomes pervasive.

If you flip back through this blog, or have seen me speak, you'll know I'm very interested in this topic, especially regarding individual privacy and the right to create, share and consume content.  Privacy is a complex thing, especially when we consider the amount of data that might be consumed by any of the companies in this space (e.g., especially in private spaces) or the rights of bystanders (e.g., who might not want to have their presence in a space recorded and monitored).

Unfortunately I couldn't be at this gathering, so I thought I'd share my thoughts on the 4 points in this Bill of Rights.  My views don't align completely with Hoover's, but we appear to share some core beliefs and motivations motivating his article. I'm grateful to him that he took the time to share his thoughts, as it forced me to confront them and put my own ideas into words.

In response to his four ideas, I would suggest four slightly different alternatives:

1. Many Realities, Available to All

2. A World Safe by Default

3. The Right to Digital Property

4. The Right to Augment

I won't assert that these cover all of the concerns and right we might consider in this brave new world, but they're a good start. Hoover did a good job touching on some of the main areas we need to consider.

Let's go through Hoover's suggestions one by one, to see where my four came from.

1. One Reality for All?

The very first (denoted in bold type face) assertion in his article is that "A shared, public reality between different AR systems must be the default experience". I agree with the motivations that Paul presents, especially the need for interoperability and the need for people to join each other in a consentually shared reality (to use a concept from Vernor Vinge's Rainbow's End). I also agree with the statement that if "sharing in AR requires effort or is harder than holding up a phone to someone next to you, it will train people not to engage".

But his motivations don't really support that conclusion he comes to, nor the related idea that content should be "public by default" (an idea that is, to be blunt, terrible). Instead, a more reasonable goal is that user's need to be able to share things, and be able to see and consume things others have shared.

I will admit up-front that I'm not as worried as some about our eventual ability to share and access content. Regardless of the fears expressed here, I don't see content being tied to platforms, and many of the big companies already recognize that locking content into one platform will undercut the potential of AR. Google and Microsoft, for example, are already building tools for cross platform anchors and spatial knowledge sharing, because they recognize that unless AR content is available on all platforms, it won't go anywhere. There's a long way to go (e.g., neither Google nor Microsoft are opening up the specification for their anchors), but we're on a good path.

There's a subtext in the arguments here (common in many abstract discussions of AR) that ignores the critical point that AR content doesn't exist in a vacuum (or in some abstract "AR cloud"). Any "AR content" or "AR layers" will necessarily be hosted, managed and tied to some sort of computing system (service, app, etc). The content will live somewhere, be managed somehow, and be accessed and interacted with by applications. There are no Pokemon in the park; they exist in some collection of servers somewhere, and are accessed through an application on your phone, and presented to you. This may sound obvious, even silly, but its important to remember.

I'm enthusiastic about the web continuing to be that basis for many of the systems we use, as it largely is for content now. I imagine there will be many AR sites on the web offering up many takes on AR layers and AR systems for consumers and enterprise users, much like the multitude of websites on the web now. There won't be a MagicVerse that everyone uses, but Magic Leap might offer the MagicVerse as a software system that people might choose to use on various platforms, or not. They might use a FacebookVerse for connecting with friends or a GoogleVerse to access some spatial version of Google Docs, along with a vast collections of other sites that make up The Metaverse. Want to see my content? Come visit the BlairVerse at (where you are reading this).

What's needed to make this happen is pretty simple, because you are already using it: web browsers. Not current 2D web browsers, but XR-capable web browsers that run on XR devices, present XR content pulled from web servers, and make both legacy 2D and new XR content available with at the characteristics that make the web so successful now: standards-based; secure; private; with users in control over how and where and when to access content.

If we want content to be accessible to everyone, and we want anyone to be able to create and share content, we need to be working to create the XR Web. All that's needed from our platforms is the ability for them to work with these sites, so that users have the choice of what platforms and systems they want to use.

Ironically, the online chat I missed was done in Mozilla's Hubs, a compelling example of the power of the web for XR. Hubs is implemented on the web platform, and can be accessed and used on (almost) any platform, from desktops and laptops to phones and tablets to head-worn displays, as long as those platforms have a capable browser.

Therefore, I'm rephrasing Hoover's first point as

1. Many Realities, Available to All

2. A World Public by Default?

As with the first point, I find myself empathizing with some of the motivations behind the assertions in this section, but feel the conclusions are unrealistic and unworkable.

Yes, we will want to be able to share what we're doing with others with a minimal amount of friction, and we may want systems that allow us to share some information about what we are doing with others, just as people choose to share with their Steam or Discord friends the music they are listening to, the videos they are watching, or the games they are playing. The keys bits of that statement, though, are "choose to share" and "with friends".

The idea of erecting virtual privacy walls, or manually controlling what is and isn't private, sound interesting in theory, but I don't think they are practical. AR researchers have explored ways of controlling what content is private and shared in collaborative AR spaces for many years (e.g., including work that happened in the lab I was a grad student in back in the 1990's[1]), and manual controls just don't work; people leave things in the wrong state, for example, exposing things that should be private. Worse, the idea of virtual walls is great in a deserted space (like the video in the article), but imagine being in a mall, or on a crowded street near Notre Dame in Paris. The clutter would be unimaginable.

At a more fundamental level, however, there is no way we want a system that automatically tells random people (nearby, or otherwise) what we are doing on our devices. Hoovers "my glasses let people know I'm in a Pokemon layer" is a privacy nightmare; sure, I might choose to share that I'm playing Pokemon GO, but probably only with friends or other people who play Pokemon GO.

In the real world, a world of stalkers and trolls and bigots and so forth, I don't want random people in the world knowing anything about what I'm doing, be it Pokemon, or exercise, or following directions to a restaurant, or swiping through profiles on some AR version of Tinder. The right to anonymity is as important as the right to privacy, and everyone should have access to the benefits of modern technology, even if they wish to remain in the shadows.

One key thing that platforms might be compelled to share, however, is when the technology in use by the wearer of the AR device affects us as bystanders. When we think back to the negative reaction to "Glassholes", it was driven by the perception that the only thing someone wearing a Google Glass was likely doing was recording video or otherwise surveilling you. If you are wearing an HMD, I don't care if you are playing Pokemon GO; I do care if you are streaming video of me to a system that detects I'm there and adds my presence to a tracking database. Beyond privacy, the individual sites and applications being used can still support connections and awareness when appropriate. I don't need to know you are chatting with a virtual friend; I might care if your friend is also a friend of mine, and our joint social media app might offer us the opportunity for a serendipitous meeting.

Future systems will need a way to share these mundane bits of information to the systems around them, but should not be sharing higher level content. That should be handled by the applications and services each of us choose to use, via the capabilities provided by those services. I may have my LinkedIn profile floating above my head, visible to anyone 1 or 2 connection steps from me on the site. Or my Facebook status, visible to those people who could see it on Facebook. Or the games I'm playing, visible to my Discord friends. Perhaps I'll create an avatar on some site that other users of that site will see instead of seeing the real me.

We need to feel safe using AR devices in public, and having others use them around us. We need to be in control. We will develop social conventions around these devices, but we shouldn't be arguing for an infrastructure that requires us to disclose anything about ourselves to others we might pass by (except perhaps the things that affect them, like if my devices is sensing and capturing digital traces of them).

The second point might be changed to be

2. A World Safe by Default

3. The Right to Digital Property?

On this, Hoover and I strongly agree; I've written and talked about the need to control what systems capture in your private space, including proposing very similar ideas to those in the article (I've even proposed an architecture in a research paper about our Argon system that included "infrastructure servers" for organizations and personal homes[2]). I won't repeat them here.

Suffice to say, I believe it's crucial that we do not allow the so-called AR Cloud startups to automatically gather and aggregate maps of our private spaces; once we start allowing companies to collect this data, regardless of what they say, "it's out there" and can't be pulled back. Witness Roomba deciding it wants to sell the maps its vacuums create of your home or Google backtracking on their pledge to keep Nest data completely separate from the rest of the data they collect.

Let's leave this point as is:

3. The Right to Digital Property

4. Freedom to Assemble?

I'm ambivalent about this point, as it's presented, as well as the use of the word "Assemble". I think the strong assertion that "people must have the right to peaceably assemble in AR public spaces, regardless of borders" is asking too much, and risks trying to impose Western values on the world.

The examples also conflate the ability to remotely visit a public space via AR with the right to use AR in a space. How would I "'land' in a new location and see others walking on the street in real time, and go have a conversation with them"? What is the infrastructure that would make that happen? Who owns it and funds it and controls and monitors it? How do we avoid the inevitable harassment, abuse and disruptions such a vision would enable?

Setting aside the "right to remotely visit a place", the underlying question could instead be thought of as "the right to create AR systems that exist in public spaces." Pokemon GO raised all kinds of issues for people about when and were it's ok to put augmentations (see "Holocaust Museum to visitors: Please stop catching Pokemon here or guidelines such as 3 Ways to Keep ‘Pokemon Go' Players out of Your Yard"), and these issues need to be addressed. As with point 3, I've written about this in the past, but my belief differs from Hoover in that I believe that individuals have the right to augment the world as they see fit, but those rights end when the content is public and available to anyone as part of the commonly-visible-to-all content of an AR system. I have the right to paint virtual content on your house under the same rules as my current ability to write about you or your house (e.g., libel, slander, hate speech, fair use, etc). Put another way: I should be allowed to cause you to appear any way I want to me, but you should have some say in how I might use a system to affect how you to appear to others.

I'd change this point to

4. The Right to Augment

So that's it, some thoughts on some "rights" people should have in future AR systems.  I'm sure my thinking will continue to evolve, and I'd love to hear what people think about these ideas.

  1. ↩︎

  2. ↩︎

Banner image from MaxPixel