XR needs ethics
Avi Bar-Zeev has been at the forefront of XR and the Metaverse for over 30 years: first launching Disney’s groundbreaking Aladdin VR ride in the early 90s, crafting Second Life’s 3D worlds and co-founding Keyhole (Google Earth) mirror-earth 3D browser around 2001. He co-invented Microsoft Hololens in 2010, helped define Amazon Echo Frames in 2015, and then contributed at Apple on undisclosed projects. Since leaving Apple, he’s consulted privately for top XR companies and investors. And he’s taken a more public stance by writing and speaking publicly on important issues like privacy, biometric sensing and mental autonomy, targeted and future experiential advertising, harassment in XR, and various open Metaverse/web3 architectures. Most recently, he is a founder and President of The XR Guild, a new non-profit membership organisation that seeks to support and educate professionals in XR on ethics and most positive outcomes for humanity. Anelia Heese from the AUREA Community team sat down with Avi for a conversation on this thing we call the “metaverse”, storytelling, how the future of XR is intertwined with the ethics of AI, and whether you should build your next startup on ChatGPT.
When we dial in, Avi’s Zoom background is a massive bookcase expanding the entire screen. During the pandemic, the “expert bookcase” has become the preferred background for applying a patina of authority to an amateurish video feed. But Avi’s bookcase is not just for show. He loves to read and write: “I always wanted to be a writer a few times throughout my career, but I’ve never been a very successful writer.” We talk about the desire to produce good, meaningful work. He laughs: “The most important discussions surrounding writing only come much later, and hence the author is often not even aware that they have published seminal work.”
Sci-fi etymology and tech
Avi has done extensive research on the origins of common terminology used in the tech industry. He studies the subsequent morphing of these original terms into different contexts with completely different meanings. In his 2022 AUREA talk, he uses etymology to explain this very moment in tech: what are we building, and what are we calling this thing we’re building?
The word “metaverse” was first coined in 1992 in the sci-fi novel, Snow Crash by Neal Stephenson. The novel depicts a dystopian future world where rich people escape into an alternative 3D, connected reality. “Some people read these books and get excited about this version of the world. I wish people actually understood that books like Snow Crash actually contain a warning of our current lived experience.”
Avi recently published a thought-provoking article on Medium titled “Coming Soon: You are the Ad” that highlights the dangers of generative AI in advertising. In April 2023, Meta CTO Andrew Bosworth announced that using generative AI for advertising will be a top priority for this year. Bosworth believes that AI can help a company make different images for different audiences.
What could possibly go wrong? Avi strongly feels that there are “real harms of targeted advertising since it could diminish our ability to control our own future.” He explains: “Currently, advertisements don’t truly know you. However, it’s not too far off that companies will be able to really understand what pushes your buttons: what makes you angry, jealous, sad, happy? There is currently no limit on how they would be able to manipulate you.”
Storytelling & AI
He says the current discourse surrounding AI taking away jobs is misplaced. “Technology has been taking jobs for centuries, and has mostly replaced it with better jobs. I don't want to have to write code. What’s really unfair, is that artists who have honed their skills have had their work fed into AI models without their consent.” However, we’ve always borrowed from somewhere, even as we shared stories around the campfire, millenia ago. “What we don’t want is technology replacing something that is core to our humanness: storytelling.”
Storytelling is at the heart of what Avi has been doing his whole career. “I had no idea you could tell stories, and make money making computer graphics!” When he saw the 1982 sci-fi movie Tron as a teenager, he was hooked. “It was amazing! I wanted to know how they did it. At the time, the only computer I had was a Commodore 64. I made a cube rotate on the screen using instructions from my maths books from high school using trigonometry.”
His original inspiration, the movie Tron, kept drawing him to work in the movie industry. “I could get a job at Pixar, but then I’d be one of 300 people working on the same movie, say, how to get this rock to look just right. It was a long ladder to get to work on a character. But I wanted to direct the whole movie straight away. I wanted to create the holodeck!” (For those who missed the Star Trek reference: The Holodeck is a fictional device which uses "holograms" to create a realistic 3D simulation of a real or imaginary setting, in which participants can freely interact with the environment as well as objects and characters, and sometimes a predefined narrative.)
“Virtual reality was popping off in the 90’s. The internet was just starting out, 1992, ‘93, and everyone called it cyberspace, which was a synonym for virtual reality back then. But in 1992, computers really sucked. If you wanted to create a cube, you had to buy an expensive program and draw all 6 sides of the cube and tediously connect them. It took me days to create simple simulations and I realised why Pixar needed 300 people to create a movie, with some people spending days editing an image of a rock. I kept wondering, how can I make this easier? You could start by creating a smaller plane, or view window, with a virtual reality headset, but they were like giant washing machines on your head and the frame rate was very low. It was terrible!” The solution was a more comfortable virtual reality environment, “which we ended up calling the Virtual Environment Theatre. Carolina Cruz-Neira coined the virtual reality interface as CAVE.”
If you ever studied Philosophy in high school, you will recognise the reference to Plato immediately. In Plato’s Cave, the philosopher uses an allegory of moving shadows in a cave to contemplate perception, reality and illusion. The CAVE model, or cave automatic virtual environment, is an immersive virtual reality environment where projectors are directed to between three and six of the walls of a room-sized cube.
Avi chuckles at this reference to this age old philosophical problem: “What is reality anyways? Every time you peel a layer, there could be another layer underneath you.” Avi and his team built their own version of the CAVE, what they called the virtual environment theatre. As a first concept, they tried to sell it as a kitchen planning tool to the home depot retailer, Lowe’s. However, it was still too impractical and expensive. A following project using the same concept seemed to crack the code. “We created this virtual hot air balloon ride through Giza, 3000 B.C. You had this joystick and could navigate through this beautiful landscape, when all the pyramids were still white, the Nile was lush. It was magnificent, but the company ran out of money.”
This failure opened an opportunity at Disney, where he could work on a top secret project: a custom-built headset. “The visuals were amazing. I have never seen anything like that. We built this ride that was based on the Aladdin movie where you can fly your magic carpet through the animated landscape. That’s when I knew: this is what I wanted to do. This is active, dynamic storytelling.”
Disney had the money to spend on these projects, where most others did not. “Nothing about this was economical - they wanted to get ahead of the game. They were right: the $20 million they spent on hardware and software back then allowed them to know how to do the things that would cost them $100 million later. They spend just enough money to try something. When it's not practical and when nobody else is doing it - that’s exactly when you should try it.”
He smiles at this thought: “Maybe my career has been about spending other people's money early on, to try things that nobody else can try.” But he reckons the cycles are getting shorter: “Back then, what we did was 20 years ahead of its time. I can’t talk about the work I’ve done for Apple in the past, but I reckon we’ve worked on stuff that was five years ahead at the time.”
Tech needs ethics
Is it significant that the companies working on these new tech frontiers are rooted in storytelling, human connection, the arts? That we are willing to spend money on projects that help us tell better stories, connect better to each other? “This is part of human culture. It's inescapable, it's important. It's not frivolous. It's the way we connect with people and it's the way we teach people about the future.”
Are you scared about the future we’re currently building? “I always had this goal of trying to help the world through storytelling. But technology is not neutral. You have to steer it towards the good or else you will get the worst by default. You can’t just build whatever is convenient and make money. Facebook is the epitome of this for me: what they do can be really harmful.
If you study civil engineering, where you might build bridges one day, you have to take a course in Ethics. If I were to become a doctor, I'd have to take the Hippocratic Oath. Even lawyers - even if we don’t think of lawyers as the most ethical people - they have industry standards. We don't have that in computer science. That’s why I started the XR Guild: to create a community that agrees that ethics is important.”
We start chatting about the current hype around ChatGPT. “People might look at language models now and think it’s a bit scary. But very soon, we can put sensors on our head and all that data can be used to undermine and exploit you in ways we cannot foresee yet. I’m very concerned for the artists whose works have been involuntarily incorporated into machine learning models against their will, without their knowledge. Some other company is making money off them. We should be paying those artists.”
According to Avi, this behaviour of OpenAI holds a clear warning for startups who are currently building new products using ChatGPT:
“Maybe you’re building a startup on ChatGPT, and that’s great when you are one of the ten best companies, who they might acquire. Maybe you believe they will pay you for your idea. But maybe they won't, because they're not paying anybody else: they're not paying the artists, the writers, all the people they got their data from.”
Dystopian Times Ahead?
He recalls an allegory to compare Big Tech to illusionists and magicians. “The illusionist’s job is to manipulate you for entertainment. But if they’re taught to manipulate you for profit, then it’s different. If you get home from a magician’s show and realise the magician stole your watch but didn’t give it back, you wouldn’t just feel tricked. You’ll feel robbed. At the moment, people don’t care because it’s still enjoyable, and it’s entertaining while it happens. But what if they turn to other emotions, like making you enraged? Maybe, as we’ve seen with the Cambridge Analytica scandal, what happens if you don’t realise you’re being manipulated?”
Avi knows that there is nuance in any conversation about ethics. But as he’s wrapping up our conversation, he says: “If there is one rule that we can establish today, the one rule that I think would do the most good, would be: AI should only ever be allowed to help you in a way that you have explicitly asked for. We have to work really hard to make sure AI is not biased, and that AI cannot be used to undermine a person’s cognitive freedom, their autonomy, and their ability to think for themselves. There has to be repercussions for companies who are planning to undermine this. It should not only be considered wrong - it should be illegal. And we all know, fines don’t work for Big Tech. It should be jail time.”