Exploring the intersection of AIGC and Web3 games: How to tell the story of Web3 games with AI?

Author: Overnight Porridge, The Path of DeFi

In the recent backdrop of the explosive growth of AI-generated content (AIGC), the Crypto market seems to be somewhat cooling down, leading to the following meme:

Regardless of whether this trend will continue, one concern I have about the current AIGC is that if it adopts a centralized approach, it will lead to the concentration of power, which can result in corruption—a future I do not wish to see.

However, it is undeniable that the influence of AI is growing stronger, with various AI models emerging, such as GPT-3, DALL·E, Stable Diffusion, Midjourney, Stable dreamfusion, GET3D, etc. From generating 2D images to videos, and then to 3D modeling, AI's capabilities are developing at an astonishing speed.

So, are AI and Crypto unrelated?

1. Examples of AI and Web3 Game Integration

In my view, they complement each other, especially concerning the super narrative of the metaverse.

Imagine a future day when we can access magical virtual worlds through AR/VR devices, where we can instantly create 2D and 3D objects from our imagination or beyond using prompts, as if we had uttered a magical incantation, and truly possess them (with data hosted on a public blockchain). We can also interact with intelligent AI NPCs in the virtual world and influence the story development of the entire game world, all supported by a completely transparent open-source infrastructure.

Image source: Generated by the Boundless Map AI tool.

In this vision, humanity will unleash infinite creativity.

It is this beautiful vision that drives my interest in the open metaverse and AI-driven gaming, and below are four projects that have currently received VC investment.

Note: The attention mentioned for the projects refers to the interest of top VC members in the projects, rather than the number of fans or Discord members.

1. AI Arena

AI Arena is an AI-driven Web3 competitive game developed by ArenaX Labs, which completed a $5 million seed round led by Paradigm and participated by Framework last year. In this game, players design and train AI-driven NFT fighting characters in a global PVP arena and let these characters battle automatically.

Image source: ai arena

Participants in the game are mainly divided into competitors and researchers, with researchers powering the game competitions by creating the best machine learning (ML) models and earning the game's native token, Neuron ($NRN), which can promote the development of AI algorithms. On the other hand, players from around the world compete to devise the best training strategies to maximize their NFTs' performance in the arena, with higher global leaderboard rankings yielding greater token rewards.

Game difficulty: Easy to start, hard to master.

Highlights: Promotes the development of AI algorithms.

Implementation difficulty: Relatively easy.

Token issuance status: Not yet issued.

Attention: Average.

2. Delysium

Delysium is a Web3 FPS game with AI elements, which completed a $4 million seed round in March this year, led by Alameda Research, Republic, and Galaxy Interactive, and a $10 million strategic round in October, led by Anthos Capital and participated by Immutable.

Image source: Delysium

In this FPS game, the biggest highlight is that AI-driven MetaBeings can hold private property and participate in the game to earn income. They are no longer characters controlled by players but AI-driven robots.

Game difficulty: Easy to start.

Highlights: AI robots are smarter than traditional computers.

Potential issues: For incentivized blockchain games, AI can also bring destructive power, making it important to design a good economic model.

Implementation difficulty: Relatively hard.

Token issuance status: Not yet issued.

Attention: Average.

3. rct.ai

rct.ai is a startup providing AI solutions for the gaming industry. In 2019, the team received millions in seed funding from the well-known startup incubator Y Combinator, followed by a $10 million Series A investment led by Makers Fund. In November 2021, rct.ai secured another $10 million A-3 round funding from Metaverse Capital, Galaxy Interactive, HashKey Capital, and others.

Image source: rct.ai

The most important product of rct.ai is Chaos Box, which is described as an AI engine that uses the Chaos Box algorithm to analyze real-time player inputs and dynamically generate NPC responses and new storylines based on deep reinforcement learning. It can support large-scale emergency behaviors of AI and NPCs in games.

Short-term highlights: Profitability through AI robots in P2E games.

Development difficulty: Relatively hard.

Attention: Poor.

4. Webaverse

Webaverse is an open metaverse project composed of various open-source software. In 2021, the team received $2 million in seed funding led by 1confirmation.

According to the project whitepaper and GitHub code repository, Webaverse provides an open-source, browser-based metaverse engine that anyone can use to build and host virtual worlds and gaming experiences. The Webaverse team is currently creating an open metaverse world called Upstreet based on this engine. In the Upstreet virtual world, players can randomly generate 2D or 3D assets using various AI models (Note: Webaverse forked and modified open-source AI models like Stable Diffusion, Stable dreamfusion, GET3D, AdelaiDepth, etc.) and mint them as NFTs on its provided Ethereum sidechain to import into the virtual world. Additionally, Upstreet will also apply AI to NPCs to enhance the gaming user experience.

Image source: Webaverse

Game difficulty: Easy to start.

Highlights: Incorporates various open-source AI models.

Potential issues: Limited funding may not match the grand project vision.

Implementation difficulty: Relatively hard.

Token issuance status: Not yet issued.

Attention: High.

2. Using AI to Tell Stories in Web3 Games

Storytelling is one of the oldest and most powerful forms of communication. When a story captures our attention and engages us, we are more likely to absorb its information and meaning.

An important point of metaverse games is that they require users and platforms to participate in the creative process together, rather than being limited to pre-planned games. In games, we have a concept of Lore, which is predetermined by game designers in traditional games and is completely predictable. However, through AI models, we can gather various inputs and generate unpredictable outputs, giving such games infinite possibilities.

Image: This is a model of a neural network; AI like GPT-3 is much more complex but has a similar structure.

This is another potential area for AI in Web3 games, aside from AI-generated content and AI NPCs.

3. Some Thoughts

Compared to the already mature field of AI image generation, the AI gaming sector involving 3D assets is still in a very early stage. Because of this, the funding amounts for related gaming projects are still relatively low.

However, it is undeniable that the market size in the gaming sector far exceeds that of the image market, and AI will be a very important trend for the future development of games.

Therefore, in my view, the potential of AI + Web3 gaming will be enormous. However, this exploratory path will inevitably encounter bumps and challenges, and the early projects mentioned in this article are likely to face failure, but their exploratory directions are worth learning from.