In the era where traffic is king, AI is engaging in fierce competition among traditional apps, vying for user traffic and market dominance.

According to a report by technology media The Information on Wednesday, American food delivery giant DoorDash has recently taken a series of technical measures to prevent OpenAI's AI agents from using its platform. DoorDash has not only deployed technical defenses to detect and block agent requests but has also explicitly stated that it prohibits any unauthorized third parties from using its services.

This is not an isolated case; many application platforms internationally are taking measures to restrict AI agent access.

The core concern of these companies is that once AI agents can seamlessly access various applications, users may bypass the native app interface and place orders or complete transactions directly through AI agents. This means that the interfaces, recommendation systems, and marketing channels painstakingly built by app developers could be undermined, losing direct contact points with users.

In other words, while this may seem like an interface issue on the surface, it fundamentally concerns profit distribution and the reconstruction of business models.

Recently, a source close to DoorDash revealed to technology media The Information: “If your service is replaced by an AI agent, you lose the ability to promote special offers, cross-sell, and control the user experience. This directly threatens customer acquisition costs and profit margins.”

The concerns of traditional app platforms are not unfounded, as there are signs that AI agents may be able to replace all possible devices in the future.

Taking the domestic large model "Liu Xiao Hu," one of the AI models, as an example, it recently upgraded its AutoGLM large model, which, according to the official statement, can theoretically perform any task that humans can do on visual electronic devices (computers, mobile phones, tablets, etc.).

Another core issue of concern for technology companies is the control over user data. When users interact with services through AI agents, who can access, analyze, and monetize this interaction data? This data is crucial for optimizing recommendation systems, personalizing services, and targeted advertising.

Once OpenAI's agents become the primary interface for users, they will gain unprecedented user intent and behavior data. This could reduce the original application platforms to mere backend "infrastructure providers," losing data sovereignty