Summary of Key Points
Recently, AI companionship products from companies such as MiHuyou (e.g., "BSide:Olivia Lin") have been discontinued due to periodic adjustments, but the AI-native game sector has become exceptionally hot: numerous startups are entering the market, and investors' interest is growing (with 70 products displayed at ChinaJoy). However, most of these products merely incorporate AI chat interfaces without any innovative gameplay. The industry faces challenges in commercializing AI interactions, such as the high cost of tokens used for these interactions. Various models are being explored, including buyouts, DLC, and hybrid monetization strategies. A few teams have creatively turned AI's limitations (e.g., "illusions") into unique gameplay elements. Overall, the industry is at an early stage similar to that of mobile games—chaotic but full of potential, waiting for a breakthrough product that can successfully combine both innovative gameplay and viable business models.
Detailed Analysis
1. Contrast between Decline and Growth
The discontinuation of AI products from companies like MiHuyou and Anuttacon (founded by Cai Haoyu) might suggest a cooling of the sector, but the opposite is true:
- Sector Heat: The AI game area at ChinaJoy was crowded with players waiting to try out games, and investors were actively engaging in discussions (although they were cautious, their interest was evident). The number of AI game investments since the end of last year has far exceeded those of the previous two years, and team valuations have also increased.
- Why the Contrast? Early products may not have provided a distinctive experience (they were just chat interfaces without substantial gameplay), but the potential of the AI technology itself is seen as promising. AI can offer interactions that traditional games cannot, such as dynamically generating storylines and personalized characters.
2. 90% of Games Are Just Using AI Concepts?
Liu Yi from Kingsoft Worldtour believes that 90% of current AI games are merely applying AI concepts:
- Manifestation of Conceptual Use: These games simply integrate an AI chat interface and change the setting to a urban or fantasy theme to create a new product, offering no different experience from regular chat apps. The gameplay would still be viable even without the AI component.
- Definition of True AI-Native Games: A true AI-native game provides experiences that traditional games cannot achieve. For example:
- AI-driven Group Adventures: Players' real-time choices create a vast number of story branches, which are impossible to pre-plan manually.
- **"Redemption Wave": AI-powered NPCs simulate phone scams, requiring players to navigate conversations based on the NPCs' personalities—this level of interactivity was previously unattainable.
In short, AI should be the core foundation of a game, not just a decorative element.
3. Commercialization Challenges: Who Will Pay for the Token Costs?
The cost structure of AI games is fundamentally different from that of traditional games:
- Traditional Games: They rely on a large number of free users to cover server costs.
- AI Games: Each interaction and character creation incur token fees, meaning the more active the players, the higher the costs.
- Solutions Being Tested:
- Buyout + DLC: The basic version of "Maggie's Garden" costs 50 yuan, which includes enough tokens to complete the main story. Hardcore players can purchase additional DLC or access APIs directly, paying for the extra tokens.
- Monthly Subscriptions: The 328-yuan "EVE" subscription covers token costs for text and image generation, but this has sparked controversy.
- Hybrid Monetization: "Babel Library" aims to create a mobile version where players can trade cards and world-building components (similar to Roblox's UGC model) to generate continuous revenue.
The industry is still searching for the best approach.
4. Turning AI’s Weaknesses into Gameplay
AI has a significant issue: its tendency to deviate from the intended storyline or worldview. Some teams have turned this into unique gameplay:
- **"Just Pet Him": Players take turns chatting with an AI, and the one who makes the AI say a specified phrase first wins. Opponents can deliberately distort the AI's perception (e.g., telling it that melons are plastic), turning these misunderstandings into strategic elements.
- **"Redemption Wave": Players act as scammers using AI-driven NPCs to simulate phone scams—a truly immersive interactive experience not found in traditional games.
These innovations transform AI’s drawbacks into distinctive features.
5. The Industry Is Like the Early Days of Mobile Games: Chaotic but Promising
Current AI games are similar to mobile games around 2010:
- Current Situation: Products are rough, with unclear definitions (anything labeled an AI game is included), and business models are controversial; many early projects will fail.
- Prospects: More and more people recognize the potential of this technology. With 70 AI games this year and possibly 200 next year, enough teams will experiment until a breakthrough emerges, similar to "Angry Birds"—a game that relies on AI, is widely accepted by users, and generates revenue.
Investors are cautious, but they see the potential: "The industry is in its infancy, which makes it exciting."
Conclusion
AI-native games are still in their early stages, with some products being mere gimmicks and facing commercialization challenges. However, there are also innovative approaches that turn AI’s limitations into valuable gameplay elements. The future winner will be the one that can make AI a core component of games while effectively managing costs. Just as touch screens defined mobile gaming, AI could define the next generation of gaming.