Bianews

Who could have foreseen that the latest hot topic in China's mobile phone industry would not be sparked by a traditional mobile phone manufacturer's Mate or Pro model, but rather by the unconventional collaboration between ByteDance's Doubao and ZTE?

Previously, 'AI mobile phones' seemed to be a claim every flagship brand could make, yet most simply integrated large AI models into voice assistants as a gimmick. In contrast, the Doubao mobile phone takes a fundamentally different approach. It first equips the device with a comprehensive set of system-level AI agent capabilities and then finds a suitable mobile phone to house them.

As a result, the discussions it has ignited extend far beyond the product itself. The crux of the debate is whether this represents a well-deserved innovation or a reckless gamble.

To answer this question, we must first delve into how it emerged and how swiftly it encountered obstacles.

Doubao Mobile Phone Suddenly Sparks Heated Debate

The weekend the Doubao mobile phone was unveiled, the tech community was suddenly thrust into the limelight.

In videos, AI seamlessly navigated between apps on the mobile phone, recognized interfaces, automatically compared prices, and placed orders—all with a fluidity that left many in awe. For the first time, many realized that mobile phones could operate autonomously, rather than always awaiting human input.

A sense of anticipation for the future swept across social platforms, as if the devices we had been using for the past decade were mere prototypes.

However, as this excitement transitioned from product launches and short videos to the hands of real users, the situation took a drastic turn.

The Doubao mobile phone frequently encountered compatibility issues with many top apps. The complex chains of actions initially demonstrated often faltered at risk control pop-ups or failed at certain buttons after real-world redesigns.

Some promotional scenarios became impossible to replicate due to UI updates, and many apps displayed clear caution towards such unauthorized, fully automated behaviors. The more aggressively the Doubao mobile phone attempted to integrate into the real ecosystem, the more chaotic its interactions with existing barriers became.

The rapid degradation of functionality turned initial excitement into a nagging question: Is this truly an AI mobile phone capable of performing tasks for me, or just a figment of imagination that only exists in demonstration videos?

For users, the ideal AI mobile phone should be a device capable of 'handling' complex processes, making daily life lighter and more natural.

Yet, the current Doubao mobile phone, while showcasing an enticing vision of cross-application automation, continues to stumble in the reality of 'whether it can be used and whether one dares to use it.'

Thus, a more pointed question arises: Is the Doubao mobile phone a revolutionary being besieged, or a predator being exposed?

Pro: Doubao Mobile Phone Attempts to Rewrite Paths and Entrances

Upon its release, the Doubao mobile phone was hailed as sparking an AI mobile phone revolution.

Luo Yonghao, who once sold Smartisan's assets to ByteDance, even took to Weibo to express his support for the Doubao mobile phone: "No technological revolution can be halted, regardless of whether the resistance comes from big or small apps."

However, the Doubao mobile phone also faced immediate backlash across the internet overnight.

The 'cross-application task completion' ability initially demonstrated quickly began to erode in the real world. Deep operations on top apps triggered constant risk control blocks, and some scenarios simply stopped responding to automation chains.

The feature list of the Doubao mobile phone did not age gracefully but was rapidly constrained by external tightening, permission restrictions, path restrictions, and tolerance reductions. The scope of its operations shrank day by day, and the capabilities it could demonstrate dwindled one by one, as if being squeezed from all sides by an invisible hand.

For over a decade in the mobile internet era, platforms have grown accustomed to designing user paths, controlling traffic entrances, and planning every step of engagement and conversion—in short, trusting themselves.

What the Doubao mobile phone does is highly audacious. It attempts to fully reclaim this path for itself, becoming the user's sole agent, standing atop the ecosystem, and using natural language to navigate the user's path on their behalf.

It creates a highly enticing future scenario: By standing with me, users can bypass the routes platforms have laid out and follow only their own commands.

In this scenario, where system-level AI can execute tasks across applications, thousands of apps are no longer the sole operational entrances but merely nodes in a service.

However, the risks following convenience are often overlooked by the excited crowd.

If it were just a smarter voice assistant, no one would specifically adjust risk control strategies, change interface structures, or reset detection thresholds for it. Precisely because it attempts to utilize underlying permissions and ignore dual authorization, it has become a target of criticism.

Con: Doubao Mobile Phone Oversteps App Boundaries and User Privacy

The current path taken by the Doubao mobile phone does not adhere to a 'standard path' widely recognized by the industry.

It achieves cross-application operations through screen reading, interface structure analysis, and simulated clicks—essentially advancing GUI automation, previously used in testing or automation tools, into a system-level capability for ordinary consumers.

This approach is not technologically novel but is unprecedented in scale and positioning. It once lurked in the gray area but is now boldly pushed to the forefront.

Network security expert Qu Zilong pointed out that once AI possesses the ability to automatically execute tasks, it means it has gained near-user-proxy permission status within the system.

It can execute, click, and input across different apps, behaviors that are difficult to distinguish from real user operations at the system level. This puts traditional risk control models based on 'who is operating and whether the frequency is abnormal' to the test.

While some view the Doubao mobile phone as a revolutionary, others see it as a predator. The con arguments mainly manifest on three levels.

The first level is breaking system and app boundaries.

Originally, apps could rely on their own UI structures, interaction flows, and risk control strategies to build a relatively closed internal order. However, the mobile phone forcibly breaks through these boundaries from the outside through visual recognition and click simulation. For platforms, they no longer possess complete control over users within their own apps but must share this power with a system-level AI.

The second level is suspected privacy infringement.

Screen reading means the AI can see everything the user sees, but users do not know how long the AI has been watching in the background, how much it has recorded, or whether it has been uploaded to the cloud. Promises in privacy agreements, such as 'no retention' and 'only local processing,' currently lack independent, verifiable audit mechanisms.

This asymmetry creates a psychological sense of 'panoramic surveillance' fear among cautious users.

In the absence of regulatory and industry consensus, should users trust tech companies' claims of not collecting user information?

The third level is altering the ecosystem's rhythm.

Before regulatory and industry standards have established boundary rules for system-level AI, the Doubao mobile phone has chosen to 'act first and see how to clean up later.'

It employs an unlicensed operational logic, plunging into a landscape where platforms, regulators, and users are all unprepared on how to coexist with a system-level AI that can 'read, click, and remember.'

The current state of the Doubao mobile phone is more of a structural contradiction. When technological capabilities rush too quickly into a space where rules have not yet matured, it is naturally seen as an overstepper.

We cannot determine whether the Doubao mobile phone has a predatory subjective motive, just as we cannot determine whether a product will always remain pure. However, the technological blade it wields has indeed extended into what are currently forbidden zones.

Our Perspective

Returning to the initial question: Is the Doubao mobile phone a revolutionary or a predator?

If we only look at its current state, it is easy to reach a simple binary judgment. However, let's consider it dialectically: it more resembles a troublemaker who has laid 'temporal dislocation' bare on the table.

Its encounters with app risk controls, bank restrictions, and various doubts prove that its sudden appearance has struck the most sensitive nerve of the AI era: where the next generation's traffic entrances and user attention will lie.

This proposition itself is not wrong, but it employs a technological path that has not been fully digested by regulatory and industry consensus, walking on the edge of red lines regarding permissions, privacy, and transparency.

However, perhaps the truly worthy discussion is no longer about the identity of the Doubao mobile phone but rather: In the era of intelligent agents, what should mobile phones be, what should apps be, and what kind of sovereignty should users possess?

When an AI can execute numerous operations on your behalf at the system level, do we still accept the cumbersome user privacy and security authorization methods with multiple platforms?

When data is no longer just statically stored but is being read, analyzed, and inferred in real-time, must privacy protection shift from 'only collecting necessary data' to 'no privacy infringement as long as the user permits'?

When users begin to get accustomed to telling an AI a single sentence to complete a complex process, do we still need to adhere to today's app-based interaction paradigm?

When technology always appears a few steps ahead of regulations, can society accept this gray period where 'technology has arrived, but rules have not been fully written'?

These questions cannot be answered by Doubao alone, and it has gradually faded into quietude.

The current controversies have forced platforms, regulators, developers, and ordinary users to acknowledge: AI mobile phones are the future, but the prerequisite is to build a secure, collaborative, and fully user-considerate ecosystem.