Summary of Key Points
This article discusses the dramatic shift in the perception of AI-generated content (AIGC) from being highly sought after to becoming a problematic liability. Initially, platforms embraced AI, believing it could efficiently produce content, reduce costs, and meet users' personalized needs. However, we have now entered a new phase where platforms are using AI to combat the very AI that created the problem. The issue is that AI-generated content, while being widely consumed, lacks quality and leads to severe homogenization, which has frustrated users to the point where they are abandoning recommendation algorithms and returning to subscription-based models, threatening the very business models that rely on these algorithms. To retain users, platforms are forced to use AI to identify and remove low-quality content, which may even serve as an opportunity for them to upgrade their services and gain a competitive advantage.
1. Platforms' Attitudes Towards AI: From Treasure to Trouble
- Initial Hopes: Platforms saw AI as a goldmine:
- Benefits: AI could generate massive amounts of content quickly, without the need to pay creators, and it could tailor content to users (for example, generating pet videos based on users' preferences), solving the problem of a shortage of high-quality content.
- Subsequent Realities: As AI-generated content became more widespread, platforms realized that users were not only not staying, but leaving due to its homogenization and lack of quality. The content became so repetitive and uninspiring that it annoyed users, leading to a decline in engagement.
2. The Flaw in AI-Generated Content: Mass Production Leads to Boredom
The core issue with AI-generated content is not its quality, but its lack of uniqueness:
- Production Logic: The focus is on efficiency; creators may post 100 pieces of content a day in the hope of a hit, without investing time in refining them (e.g., using predefined templates or simple story structures).
- User Experience: Users become disillusioned when they find that, despite the attractive headlines, the content consists of standard, repetitive phrases (even in the free versions), wasting their time.
3. Users' Reaction: Trusting Only What They Follow
Fed up with AI-generated spam, users started to boycott it:
- Abandoning Algorithms: They stopped relying on algorithmic recommendations and returned to following creators they trusted, either because their content was not AI-generated or because the creators used AI effectively.
- The Disaster for Platforms: Recommendation algorithms are crucial for retaining users. If users switch to subscription-based content, platforms risk becoming mere repositories of low-quality content, unable to generate revenue.
4. Platforms' Countermeasures: Using AI to Combat AI
Platforms are using AI to filter out low-quality content:
- The Purpose: To prevent users from escaping the "information cocoon" created by AI, which is essential for generating revenue. However, this process also involves cleaning the platform from AI-generated junk.
- The Strategy: They let AI-generated content flood the market first, then use their own algorithms to filter it, driving lower-quality creators to other platforms. Once their platforms are cleaner, they can attract users from their competitors, essentially achieving a form of "business upgrading."
5. A Dark Conspiracy Theory: Could AI Homogenization Be Intentional?
The author speculates that the initial uniformity of AI-generated content might have been a strategic move by leading platforms:
- The Plan: They intentionally flooded the market with low-quality content to then use their anti-AI algorithms to target and eliminate it.
- The Outcome: Lower-quality creators were forced to move to smaller platforms, which were overwhelmed by more spam, leading users to the cleaner, higher-quality content offered by the leading platforms.
- The Final Goal: The leading platforms aim to create a market dominated by high-quality AI-generated content, achieving a competitive advantage.
It's indeed a fascinating cycle: platforms first use AI to produce spam, then use it to clean it up, and finally use that same technology to attract users. The underlying message is clear: users don't care if the content is AI-generated or not; they just want the content to be meaningful and not annoying.
This article provides a thorough analysis of the "involution" of AI-generated content and the commercial strategies of platforms, highlighting that users' real concern is not whether the content is AI-generated or not, but whether it is useful and respectful to them. All platform actions are driven by the need to retain users and maintain their profitability.