Behind the AI Spring, the Metaverse Quietly Takes Shape
Behind the AI Spring, the Metaverse Quietly Takes Shape
Awareness that AI is really going to have a huge impact has entered the mainstream with ChatGPT (though few people outside of tech circles could probably explain why it’s called GPT). An article in the Wall Street Journal (The Metaverse is Quickly Turning into the Meh-taverse) emphasized that the Spring earnings call from Meta only mentioned the word “metaverse” 7 times while the term “AI” was mentioned 28 times. The article, immediately, noted that “Meta didn’t respond to a request for comment”. The metaverse hype has been replaced by the AI hype. That’s a good thing.
There’s an adage that “AI is what doesn’t yet work”. Likewise, the metaverse is seen by many as a vague concept that’s somehow immersive, somehow virtual reality, and that will never happen. The growth of the metaverse is dependent on generative AI.
Interactivity in an immersive virtual environment goes much further than the trivial interactions we have with websites where we click a button and something happens.
Imagine that you are in an immersive virtual environment that’s an adaptation of a literary or cinematic work. Movies are scripted. Metaverse experiences will be based around a storyline but the specific actions and events will differ from player to player based on the player’s responses. In games, the dialog of non-playable characters (NPCs) are scripted. In metaverse experiences, NPCs will be generated by AI based on actions and dialog of the player.
ChatGPT superbly demonstrates the capabilities of conversational AI powered by a large language model. The conversant aspect of ChatGPT is an enhanced feature grafted over the AI language model. Chatbots are nothing new though their dialog interactions have been limited, often by design to avoid ethical conundrums.
As more people experience conversational AI via ChatGPT and other AI-powered chatbots, more people will imagine creative ideas that will be expressed in virtual environments. Chatbots (a form of virtual characters) can be trained to respond in the style of a particular character. ChatGPT exhibits a polite and pleasant demeanor. Yet, it’s possible to code other behaviors into conversational AI.
Most corporate AI offerings will be limited, under the banner of Responsible AI, as to the extent and nature of dialog interactions. We’ll see a massive number of corporatized characters (like Disney). But we’ll also find indie creators utilizing open source AI models for less savory characters. The NSFW virtual environment will stay under the radar but artistic creators will craft virtual encounters where coarse dialog and interactions are integral to the experience.
Not everyone wants to drop into a Marvel environment. Some of us may find more entertainment descending the passageway into a virtual world more typical of a Gaspar Noé film.
For any type of content, generative AI enables more natural interactions with imaginary avatars.
AI also drives computer vision, which is an essential component of VR devices for detecting hand gestures and body movement. Refinements in AI-powered computer vision leads to more sophisticated user interactions. As devices learn more spatial awareness, which is enabled entirely by AI, experiences can be made to respond more creatively based on the player’s location in virtual space.
In the years ahead, we’ll likely stop using the terms AI and metaverse so often. The terms Internet and Web are still around though few know the distinction between the two. The Wall Street Journal business analysts do not parse earnings calls for how often Zuckerberg mentions the “Web”. That technology has found its way into the fabric of our lives. In 2043, instead of talking about AI and metaverse, we’ll be talking about experiences and environments.
