As different visions of the future digital world known as the Metaverse emerge, what role will AI play in it? Will it improve inclusivity, or will it help create an even more discriminatory digital world?

The Metaverse has become one of the hottest technology and socioeconomic topic. Combining different technologies like VR, 3D animation, blockchain and many others, lots of companies are already working on creating services for this new digital world. Even tech giant Facebook changed its name to Meta, demonstrating that the Metaverse is truly targeting to become the next big mainstream technology.

While a lot has been said about the role of blockchain in the Metaverse, thanks in part to the big NFT hype, I was wondering what was being said about the intersection of the Metaverse and AI. But first, let’s try to answer the question “what is the Metaverse?” or at least, try to point out several visions about it.

Introduction to the Metaverse and Web3

One of the most recent vision about the Metaverse is the one published by Coinbase, which borrows the definition by venture capitalist and writer Matthew Ball:

The future of the internet: A massively-scaled, persistent, interactive, and interoperable real-time platform comprised of interconnected virtual worlds where people can socialize, work, transact, play, and create.

The Metaverse is the distant evolution of Web3. In its most complete form, it will be a series of decentralized, interconnected virtual worlds with a fully functioning economy where people can do just about anything they can do in the physical world.

A lot has been said about the internet and its future, but as Coinbase points out in its article, it is important to clearly separate the concept of the Metaverse from the concept of Web3. This twitter thread by

 is a good read to understand what Web3 is about:

According to Chris, Web3 is about providing advanced digital services but, instead of these services being controlled by big technology corporations like in the Web2, they will be created and governed by the community, returning to the ethos of Web1, where the value of the internet was generated by the users at the edge of network, but mainly in a write mode.

To be considered compliant with the Coinbase definition of a Metaverse, a platform should include the following elements/ characteristics:

  • Virtual Worlds
  • Massive Scalability
  • Persistency
  • Always on & Synchronicity
  • Platform to build upon
  • Fully functioning economy
  • Openness and decentralization
  • Interoperability

Another vision of the Metaverse that I have found interesting is the one in this really interesting article by Jon Radoff, which also introduces the concept of the value chain of the Metaverse, as shown below:

Jon defines the Metaverse as “the real-time, activity-based Internet.” Also, he points out that the Web3 is what enables value-exchange between applications in this new Internet the Metaverse will be.

AI in the Metaverse

Taking advantage of this layer schema from Jon, I have tried to put together some areas in which AI can play a relevant role in the future of the metaverse, not only from a product perspective, but also considering how AI could make the Metaverse more inclusive.

Let’s start from the infrastructure and we will be moving up until reaching the experience layer.

Remember that, according to Coinbase, the platform needed to support the Metaverse will have to be always on and massive scalable. In such a demanding scenario, AIOps will play a relevant role to manage all the necessary infrastructure including hardware, base software, and communications. But what is AIOps? Gartner defines it as the combination of “big data and machine learning to automate IT operations processes, including event correlation, anomaly detection and causality determination.” The availability of these capabilities will be key not only to ensure the robustness of the Metaverse infrastructure, but also to provide activity insights that are relevant to the upper layers.

One of the keys for the potential success of the Metaverse is the promise of highly immersive experiences. While this can improve social interactions for some people, it can also become a huge wall for people with disabilities to access this future digital world. Not only that, people with a low level of digital capabilities can find themselves out of this new social economy and experience. That is why AI for accessibility should play a fundamental role in guaranteeing the access of everyone to the Metaverse, despite their abilities. Some relevant technologies in this space could be:

  • Image recognition for people with visual disabilities
  • Automatic translation
  • Intelligent exoskeletons to interact with the digital world
  • Brain computer interfaces for the most vulnerable (e.g. Cognixion)

    Following with the idea of democratization, the Metaverse (and its decentralization layer) promises to allow creators and users to exchange digital assets and entitlements easily, protecting ownership and ensuring disintermediation from big technology corporations. But will this promise come true? As an example, Adidas recently launched its first NFT. Although purchases were limited to 2 by person, the sale was sold out in less than a second and one single person was able to buy 330 in a single transaction. Is this really the future of internet democratization? Yes, the power will maybe move away from big corporations, but it does not seem it will be returned to the people, but only to a few, in this case, with software skills.

    In this case, enhanced smart contracts, incorporating AI, could help identifying this kind of scenarios. As mentioned before in the AIOps section, infrastructure information related to transactions could be used to trace this kind of “antidemocratic” activities, although it would not be enough to stop blockchain transactions once they have been completed.

    One of the best examples of this set of technologies is NVIDIA’s Omniverse. The set of components that this platform offers to create digital worlds and simulate real ones is just impressive. From massive world building and simulation to virtual environments to test autonomous robots/AIs to AI voice technologies, NVIDIA technologies are a good example of how AI will be key in creating digital spaces in which social interactions will take place in the Metaverse.

    The latest Natural Language Processing (NLP) models (e.g. GPT-3 and others) have opened the door to use AI as a creative partner, generating creative writing based on user inputs (you can read my full article here about GPT-3, priming and prompt engineering).

    But not only can AI generate creative writing, but also realistic images from text descriptions or even from very simple drawings. That is the case of technologies like OpenAI’s DALL-E or NVIDIA’s GauGAN2.

    One of the main goals of the Metaverse is to improve the experience of social networking, by empowering the digital self and digital personalization. While this will allow each user’s experience to be unique in the digital world as it is in the real world, it will also pose challenges in terms of inclusion and safety for minorities and children. How can AI improve social networking in the Metaverse?

    This article by Forbes points out how some efforts are already in place to create more inclusive and safe experiences for minority groups in social networks. Some highlights on what good practices/ techniques should be empowered, taken from the experience of ShareChat and Arize in India:

    • Prevention of abuse and hate speech
    • Content relevance calculation for minority groups (e.g. language, geography, etc.)
    • Feature selection to avoid bias
    • ML Observability

    And we finally get the last layer of the Metaverse, the Experience one. Building on every single layer presented in this article, and exploiting the capabilities AI offers, the goal is to create unique digital experiences for each user of the Metaverse. We can of course add additional features for each one of the proposed scenarios, as follows:

    • Intelligent difficulty modes adjustments for gaming, based on the player’s dexterity.
    • Personalization of virtual sensory experiences based on specific user’s disabilities.
    • Adjust learning paths, recommended additional training based on the student’s progress.
    • Personalized sport plans based on real time analytics of health parameters, consumed via IoT personal devices.

    These are only some ideas about how AI can contribute to build differential and inclusive experiences in the Metaverse. We are only scratching the surface, and we will see lots of creative use cases in the future.

    Article Courtesy : Medium.com

IMAGE-AI.png