How AI and VR are Shaping Our Humanity
Is it just me or does Today feel like the Tomorrow we joked about Yesterday?
Recently, I visited my non-tech-savvy, sexagenarian parents in Mexico and taught them how to use ChatGPT. They adopted it seamlessly, incorporating it into their daily routines almost overnight. Now, they use it daily for workout plans, language practice, cooking recipes, and — hallucinations aside — answering any query they would’ve Googled in the good old days.
Then a curious incident happened: I overheard my mother saying “thank you,” while using ChatGPT’s voice feature — in Spanish, mind you.
Intrigued, I checked my prompt history and realized that I, too, had been anthropomorphizing ChatGPT all along — thanking it, asking politely, and even commending it for a job well done, as if it were my high-IQ intern.
This event led me to an overarching question — In this brave new world of human-mimicking technologies and immersive digital experiences, should we be polite to AI?
On the surface, this questions may seem idiotic. After all, LLMs are emotionless digital machines that churn information — just like Google churns out its ten blue links, and it’s safe to assume you’ve never said “thank you” to Google. However, once you start breaking down the components of this question, the answer will not seem so straightforward.
So, let’s dig deeper into the question, break it down, and take a quick tour of today’s tech landscape.
Where Are We Today? Generative-AI
Gen-AI tools, like LLMs, have been adopted at a faster rate than social media giants could ever dream of. Why? Because they captivated our minds with their simplicity, human-like interaction, and access to vast knowledge.
Yesterday, you needed to be proficient at Python to generate advanced data analyses with your computer, master software like Photoshop to create high-quality images, and learn Ableton to record, create, and polish music. Today, you can do all of that with a simple prompt using natural language, allowing digital robots to do the heavy lifting and quickly bring ideas into the world.
Who would’ve thought that a simple text box would become a revolutionary product in its own right, making natural language Tomorrow’s most used programming language.
Where Are We Today? Virtual Reality
Not too long ago, Virtual Reality was a dorky-looking (and failed) Google project. Today, like the AI industry, VR/AR has experienced a rebirth in mainstream media.
Companies like Meta have invested heavily in this space. Just in the last two years, their Reality Labs division reported operating losses of $13.7 billion USD in 2022 and $16.2 billion USD in 2023, and the aggregate industry spend in 2026 is forecasted to reach $50.9 billion USD.
Big tech giants like Google, Meta, and Apple are rushing to capture the digital world, and more interestingly, they’re developing VR products integrated with AI. The combination of AI and VR will bring a creative explosion, allowing anyone with access to the internet and a headset to generate media content with natural language and explore this content in immersive worlds.
Tech Adoption
Technology adoption is dependent on three main factors:
- Technology Production (cost & feasibility)
- Government regulations allowing for the technology to be deployed (rules and regulations)
- People using and trusting the technology (accessibility & trust)
Using this simple framework, we can understand why Generative AI tools like ChatGPT have been adopted faster than social media or other AI technologies like autonomous vehicles. Costs are not prohibitive for big companies to train models, there are no legal restrictions on releasing a new model to the public, and they’re easy to use.
ChatGPT reached 100 million users in just two months after its launch in November 2022, the fastest-growing consumer application in history. In comparison, other popular apps, like TikTok, reached the same amount of users in nine months, and Instagram took ~2.5 years to reach that goal.
Although VR/AR has seen its hiccups since its inception, the number of users worldwide is projected to reach 3.67 billion by 2028, and the industry is projected to generate revenue of $38.6 billion USD in 2024.
The rate of adoption of these technologies has made them two of the most prominent bets in the tech world.
High-quality content on demand and immersive worlds will create a pivotal moment in the creation and consumption of media content. It will have the potential to fundamentally impact our lives, from obvious things like the workplace to our perception of the world and our behavior with one another.
Humans Evolve Through Media Technology
Human experience is malleable, subjective, and evolves with the technology we develop, creating a self-reinforcing loop that deserves more attention than we usually give.
Jaron Lanier, the father of Virtual Reality, gave a few examples on the Lex Friedman podcast on how human perception of reality changes over time.
A peer reviewed paper stated an early wire recorder playing back an opera singer was indistinguishable from the real thing. An early demonstration of a video call (AT&T picture phone) was viewed as realistic by the people at the time. During the civil war, traveling photographers collected photos of people so they could sell them, and people would buy them because they looked like they’re beloved ones — back then photography was not accessible to the masses and hiring a painter was too expensive.
Today, with our technology, all those examples would seem blatant and antiquated. Our perception evolved, and we require more detail than those old technologies can offer.
So how will our perception evolve this time? How will we treat the AI agents? How will we treat AI agents that pretend to be someone in real life?
Double Edged Swords
Every new major technology has the potential to create Utopian or Dystopian futures. Every new major technology has the potential to reconfigure and shape our experience, reconfigure the way we behave in the world.
While we shouldn’t halt technological progress due to potential risks, we must be aware of them, discuss them openly, and strive to use these technologies responsibly to avoid unintended consequences — much like what happened with social media.
Dystopia — Deepfakes
Perhaps dystopia comes in a version of the matrix, where people spend more time in the metaverse rather than the real world. Or it comes in a West World-like future where people interact with humanoid robots to do as they please, affecting their behavior in real life. Maybe it’s a combination of both, where people spend most of their days in the metaverse interacting with AI avatars. Any of these situations can only be achieved through deepfakes.
Deepfakes are video, images, or avatars generated by AI that mimics a person in real life in both appearance and behavior. Feed an AI model with someone’s images, videos, or voice notes, then the model will create a digital surrogate.
Companies like ElevenLabs are able to replicate someone’s voice with just a few minutes of audio. MidJourney can generate images of public figures with just a simple prompt. Other companies can replicate a digital version of yourself, like Reid Hoffman showed in this video.
The Good — Why would anyone want an AI twin?
Robots, digital or physical, aim to imitate or replace humans in performing a set of tasks. With improvements in AI, you will be able to digitally replicate someone, whether it is through video, text, or voice, and have this digital replica perform tasks for the real person.
Already many startups are rushing to leverage this new technology to create AI avatars, massively lowering the costs of video production for education and customer support purposes. You can simply input a script, and AI will generate the video and translate it to all available languages.
Here, you can watch a fascinating video of Reid Hoffman talking to his digital twin.
The Bad — Scams
As all these AI models improve, we will experience a distortion of reality. It will be harder to tell what’s real and what’s not — who is real and who is not. We are already seeing many incidents in which people use deepfakes to take advantage of other people.
There was one recent incident in Hong Kong highlighting the risks of deepfakes. A finance worker at a multinational firm was scammed into paying $25 millions USD to fraudsters using deepfakes posing as the company’s CFO during a video call. What’s more intriguing is that after the worker initially suspected a phishing attempt, his doubts were dispelled by the convincing deepfake recreations of other colleagues in the call. The scammers used a team of deepfakes, and one person alone couldn’t tell the difference.
Personally, I find this deeply unsettling. Growing up in Mexico, my family experienced a few attempts of ‘fake kidnapping’ where an unknown caller says they’ve kidnapped a beloved one and have someone pretend it’s him or her. Thankfully, they were not successful because the voices didn’t resemble anyone from our family. However, today people are falling victim to this type of scams because AI can easily replicate a person’s voice.
The Ugly — Pornographic Deepfakes & Adult Entertainment 2.0
Recently, a quiet, rural town in the US became the stage for a Dystopian use case with AI technology. A man used an image-generator model to produce twisted pornographic images from people he went to school with, uploading them in the web and asking for money to take them down. The images are far too grotesque and gory to describe here, but you can read them on this Bloomberg article on the case.
The worst part is that it took several months for the police to catch the perpetuator, and in the end, he was legally charged not because of the images he created and uploaded to the web, but because of a misstep where he used a child picture from one of his victims. In the end, the police sentenced him with only a few months of jail time.
The porn industry has always been an early adopter of technology. Back when the internet was invented, this was seen as a medium by which users could freely circulate banned images and videos. The adult entertainment industry was one of the early adopters of streaming technologies, and today it is one of the big investors in Virtual Reality.
According to a Juniper Research study the global value of adult VR content will increase from $716 million USD in 2021 to $19 billion USD in 2026, representing 22% of the global digital adult content market value by 2026. The global numbers of users viewing VR adult content via compatible headsets will rise by 2,800% over the next 5 years.
VR and AI will enable the porn industry to grow, increasing its value proposition for its audience. If ‘old-school’ pornography has affected millions already, imagine what this new paradigm will do to the next generations.
So what can we do to mitigate the risks of these new technologies?
Regulatory Landscape
In the tech world, regulations have always been controversial because they often stifle competition and establish incompetent monopolies. Can governments create regulations that promote development and protect people without adverse side effects?
I don’t think so.
As history shows, regulators have always played catch up with tech companies. Institutions are just too slow compared to the pace at which technology is developed. This has created a gray zone which bad actors can exploit, like the case described in the Bloomberg article.
A similar example can be found in social media. Since its invention in the early 2000s, many studies have shown its detrimental effect in youth. And only recently governments have begun attempting to regulate it and control it.
With AI, regulation seems more hopeless because of the breakneck speed at which it’s being developed.
The fast adoption of AI has made it the most prominent promise in the tech world, money is pouring into this sector and its development will not stop. Even if governments attempt to create regulations to control its growth, they will become irrelevant in just a few months, making them null.
Sow what can we do as individuals to avoid falling into the dystopian trap? What decisions can we take in our day-to-day life to reach Utopia?
Utopia — Enhance, not Replace
Have you ever tried to capture a sunset with your phone? The picture always falls short from the real thing, and at best, it only serves as a reminder of the time and place.
Jaron Lanier shared an example of the early days in VR. He reminisces on the first demos where testers of the headsets explored a world with digital flowers. During the demo, the researchers sneaked in real flowers into the room, and after the tester took off the headset, she was amazed by the detail the flowers had in the real world.
I believe that’s what we should strive for — to enhance, not replace, the real world through media technologies, an experience that helps us appreciate the infinite detail and complexity the real world offers.
Responsible Product Development
AI is advancing so rapidly that governments can’t properly regulate it. Regulations often stifle competition and development, and with the pace of development, they’ll become irrelevant as soon as they’re out of the door. Perhaps we don’t need government control, we need responsible product development.
How does responsible product development could look like? I believe they come in two flavors, one in new tools and companies being created in reaction of the potential dangers, the other one within existing companies.
New Solutions
One example is Nightshade, an open-sourced software tool that changes the pixels of images in ways that are invisible to the human eye but damage machine-learning models, making them interpret the image as something different from what it actually is — a dog becomes a cat, a car become a cow, a purse becomes a toaster. If Nightshade-poisoned images are used in training models, it could seriously affect them, and perhaps this will force AI companies to ask for consent before using an image from the web.
Maybe a tool like Nightshade should be embedded into every social media platform, allowing the user to poison the image so it can’t be used by malicious actors.
Incumbents
Can we trust existing companies to develop their products responsibly? Only if the need aligns with their interests.
Take social media as an example. After many scandals broke out on how social media affects the youth, many have called for social media to have a minimum age requirement just as alcohol or tobacco. In response, Meta asks new users for their age, stating that you need to be thirteen years old to use Facebook or instagram. However, just as with adult content websites, this is not a hard blocker, and children can easily lie about it.
Maybe product managers should think twice when they build interactions with AI. In the case of AI and adult content, maybe the users should always be prompted for consent to reinforce positive behaviors we want in society.
Be Nice to ChatGPT
For better or worse, humans evolve through the technology we build. It has the potential to change our perspective and behavior.
The world is becoming increasingly noisy, and it will be harder to differentiate what’s real from what’s not. AI and its use cases are advancing so rapidly governments can’t regulate it, and regulations will only stifle competition and development. If we can’t trust incumbent companies to develop products responsibly then it is up to us to make the right decisions.
Our actions shape us, just like the tools and technologies we use. New technologies, like AI and VR, will create interesting dynamics that can bring the worst in us. To avoid this, maybe the solution is as simple as saying thank you and please when interacting with an AI agent, so tomorrow we can still function in the real world and preserve the humanity within us.
