The hype around the metaverse is here. But our real world, enriched by artificial intelligence, can equally be an alternative to this virtual second world.
"Life punishes those who are late," goes the alleged quote by former Soviet President Mikhail Gorbachev from 1989. A sentence that could well give Vladimir Putin food for thought right now. But the opposite may also be true: Life punishes those who arrive too early — for example, in the form of far-reaching rejection by customers or target groups.
This happened to Google just under ten years ago when the technology group attempted to launch "Google Glass," the first pair of glasses connected to the Internet. Even before the market launch could get fully underway, the product turned out to be a complete flop. Data privacy activists fought against the introduction of the glasses, which allowed wearers to spy on their surroundings without the other person noticing. Consequently, people coined the term "Glasshole" for anyone who wore the glasses.
At the very end of its annual developer conference I/O, Google now presented another pair of glasses. The new model does not yet have a name, apparently no camera, and looks like a standard pair of glasses. But it can translate conversations simultaneously into other languages. In a demo video, a woman is talking to her daughter, one speaking Mandarin and the other English. And they understand each other.
Surprising as this might be, there is an exciting strategy behind it. The revival of the glasses indicates how Google sees the future of human-machine communication: as a permanent exchange that makes life easier.
In technical terms, this is called "Ambient Computing." With the help of artificial intelligence, smart agents in glasses, lamps, loudspeakers, and even entire lighting, heating, and energy supply systems communicate seamlessly with the people around them. And we interact smoothly with the machines that enrich our lives with information, translation, light, and music.
This vision is the future of Google's most important line of business – Internet search. On the one hand, the input mask into which we type the search term today will change. It will eventually be replaced by an AI virtual companion that is available around the clock to respond to our needs and desires without us having to speak or type them in. Looks, movements, and behavior, all become part of a calculation for perfectly personalized service as to what any individual might need at any given moment.
Secondly, this "environment search" integrated into everyday life will change the consumer world. For example, when looking for a dress to wear to a wedding party, I can use text, speech, or images to describe how I want the dress to make me feel or the vibe I want to give off at the party.
Google will then make me an offer that I can change and adjust until the dress suits me emotionally. Size or cut is irrelevant at first because only then will the dress be produced based on my preferred styles and body size.
The more advanced these AI language models become, the easier and more intuitive it becomes for humans
So this is about much more than just internet search. It is about the personalized creation and production of new products that enriched intelligent search has created for us –– from Haute Couture to Haute Creature.
If this all sounds like a scene from the 2013 Hollywood movie „Her," well, that's exactly what it is. It will be not only Joaquin Phoenix talking to his operating system (narrated by Scarlett Johansson, whose voice makes us fall in love with it). We can do that, too.
What came across as pure technical science fiction in 2013 is now becoming a realistic possibility. The rapid development of Natural Language Processing –– the analysis of natural language –– as an element of Machine Learning in the context of AI makes this possible. Google runs this variant of Machine Learning directly on its devices via the Tensor Flow chip, whether smartphone, Google Home, or the most recent glasses. The more advanced these AI language models become, the easier and more intuitive it will be for humans to interact with them. We will be talking to the virtual agents as if we were talking to a good friend — just as "Her" predicted we would.
Ambient Computing: Alternative Model to the Metaverse
Google's research division Deep Mind has recently presented an example of a "generalist agent" in a scientific paper that uses large language models to perform true miracles. Said agent, called "GATO," can play computer games, describe pictures, chat, or control a robotic arm to stack building blocks. This is revolutionary. Human-machine communication becomes flexible. All question forms (text, image, video, gestures) can be translated into all response forms.
For those who find their hair standing on end in the face of such prospects, a comparison with the current hype — the metaverse –– might be helpful. Ambient computing could become the gateway to an AR world, a world enriched by augmented reality, and thus an alternative model to the metaverse.
What would we rather do in the future: enrich our actual world with artificial intelligence to better navigate it and fulfill our desires? Or spend a growing part of our lives in a virtual second world, while our real surroundings wither away? One word seemed to be missing at this year's Google I/O conference: the metaverse.
Cover: Getty Images