Why generative AI is exciting for the Metaverse

It took 16 years for mobile phones to reach 100 million users worldwide. TikTok grew faster, amassing the same number of users in nine months. But when ChatGPT reached that milestone just two months after launch, the world certainly took notice. It reignited the public’s interest in artificial intelligence (AI), and how it could transform our lives. Meanwhile, the term ‘generative AI’ – the technology behind ChatGPT – got catapulted into the mainstream.

We regard AI as one of the four key technologies that can make the Metaverse possible.¹ The recent generative models have now brought us to an inflection point. These technologies can create original content, such as text, images, or code, in response to users’ prompts. In our view, they have the potential to boost productivity for large-scale virtual world creators by writing code and creating graphics; they can even help them with storytelling.

ChatGPT, for example, can generate code when prompted. Microsoft’s GitHub has a ‘Copilot’ feature that can autofill suggested lines of code. Meanwhile, NVIDIA has announced a text-to-3D content creation tool dubbed ‘Magic3D’ that can generate 3D models from short text prompts. This could be the precursor for the creation of immersive 3D graphics from user prompts. But generative AI could do so much more than just improving productivity alone.
 

How AI can engage users in virtual worlds

If you’ve spent time playing video games, you likely know about non-playable characters (NPCs). They are the characters in the game that you don’t control. Their actions are repetitive, and they have limited dialogue. A logical next question for many might be, “But what if generative AI could breathe some life into tomorrow’s NPCs?” This may sound a lot like the premise behind the 2021 comedy flick “Free Guy” starring Ryan Reynolds. But, in the real world, companies such as NetEase and NVIDIA have already started working on making AI-powered NPCs a reality.

Generative AI tools make it easier to create user-generated content without needing to know how to code. This is really useful for operators of Metaverse experiences, such as gaming companies. They can populate their virtual worlds, and keep users interested and engaged. The more time users spend on their platform, the more opportunities there are for them to earn additional advertising revenue.
 

Beyond platforms: opportunities across the Metaverse Value Chain

All the generative AI models are computationally expensive or complex. Training them and generating quick responses require a significant amount of computing power and time. In our view, this will likely lead to a rise in demand for cutting-edge CPUs, GPUs, and memory, which could favour chip designers and makers.

It could also be of advantage to cloud service providers that provide computing power to AI model owners and operators. The need to transmit data could also lead to higher demand for reliable fibre and mobile networks that are offered by telecom operators. 

AI opens up opportunities in other areas, too. A big reason why Meta has channelled resources into AI is to improve its advertising tech. When Apple launched its App Tracking Transparency feature in 2020, it shook the world of advertising. Without data access, companies that focused on targeted advertising were held back. They couldn’t target as effectively, which caused a drop in revenues they made from ads.

AI is a vast area that many businesses could use help with. This might be good news for companies like IT consulting companies. One large multinational computer software company is adding AI to their design tools to make them even better. Music corporations could also use AI to keep track of how their songs are used on social media, which could help them increase their revenues from ads.
 

Areas of concern about AI

We’re still in the early phase of generative AI, and it’s obvious that the models need to be better. For instance, ChatGPT doesn’t say where its information comes from, which could lead to copyright problems. It can also give users incorrect information, so this needs to be monitored.

For now, it’s safe to say that AI is raising more questions than answers. Society must address far-reaching issues, such as consent, the pervasiveness of AI models, and the biases inherent in the algorithms. From an investment standpoint, ethics are likely to be an issue for ESG investors when investing in companies using AI.
 

Winning the AI race

We think AI is more likely to favour big tech companies than undermine them. These big companies have way more data than even the most innovative start-ups.

There also appears to be an inexhaustible need for super-fast chips to handle the growing data needs to deliver better AI performance. Over time, the cost of these will likely go down, and we can look forward to new and exciting user experiences, as well as a clearer path towards monetisation.