Revelations from the Doubao Controversy: Whose Cheese is the AI Phone Taking?

On December 14th, in response to the online rumor that "regulatory authorities summoned ByteDance due to network security, data security, and competition issues," an informed source told Wandian Research that the information was false.

At the end of 2025, the large AI model "Doubao Mobile Assistant" under ByteDance collaborated with ZTE to launch the Nubia M153 engineering machine (hereinafter referred to as the "Doubao AI Phone"), sparking heated discussions in the technology field.

The "Doubao AI Phone" is not just a new smartphone but also a landmark signal for AI products. It marks the transformation of large model research and development from the virtual "cloud brain" to the "personal intelligent brain" that grasps real - time information of consumers.

From ChatGPT triggering a content creation revolution, to Midjourney reshaping visual expression methods, and now to the Doubao AI Phone opening up a new scenario for terminal intelligence, large models are reshaping the underlying logic of the business world with unprecedented depth and breadth.

Why are consumers willing to pay for "Doubao"?

For many ordinary investors, it's hard to understand why the "Doubao AI Phone" suddenly became popular. If you've recently come across news like "The second - hand price of the Doubao Phone has increased" or "The first batch of 30,000 units was sold out in seconds," don't think it's just a marketing tactic.

Besides the psychology of curiosity and the desire to try new things, the key reason why consumers are flocking to the "Doubao AI Phone" is that it is not just an AI phone but also an intelligent assistant that can free your hands.

Currently, most AI phones still stay at the level of dialogue interaction: you input commands, and it replies with information. At most, it can assist in writing copy, image processing, or taking meeting minutes.

In contrast, the positioning of the "Doubao AI Phone" is more groundbreaking: the built - in Doubao Mobile Assistant can directly take over phone operations. Just say "Doubao, Doubao, help me xxxx," and it can perform tasks like a real person.

For example, if you instruct it to "Search for a certain UP master on Bilibili and like each video," it will automatically enter the page and complete the operations one by one; when you need to buy electronic products, it can actively open the recommendation platform, compare the pros and cons of products, and provide purchase suggestions. The whole process is like having a real - life assistant browsing and analyzing on your behalf.

In addition, from booking tickets and hotels, generating travel guides, to taking over the movement and operations in mobile games, Doubao has demonstrated not only dialogue ability but also real "execution ability." This is not achieved by piling up parameters or modifying the interface but marks a key leap for AI from "responding" to "executing."

The core feature of the "Doubao AI Phone" lies in its deeply integrated large model capabilities. Users can directly converse with the device through natural language without opening specific applications, achieving seamless connection of information acquisition, task processing, and creative assistance, and promoting the evolution of human - machine interaction from the "command - response" mode to the "dialogue - collaboration" mode.

What's more noteworthy is that the personalized recommendation system of the "Doubao AI Phone" can provide highly customized service suggestions based on the user's conversation history, behavior patterns, and real - time context. This marks the upgrade of personalized service from simple matching based on tags to dynamic adaptation based on in - depth understanding.

In this regard, relevant industry observers told Wandian Research that the "Doubao AI Phone" can understand the consumer's context rather than just "executing commands," solving the pain point of passive response caused by function piling in AI products. This breakthrough forces the industry to rethink product positioning. Currently, several mobile phone manufacturers have accelerated the layout of edge - side AI, and a new era of mobile phones centered on "intelligence" is opening up.

The appearance of the "Doubao AI Phone" shows the market not only a technological iteration of an AI product but also a bold exploration and cross - border integration of the application boundaries of new technologies. In this transformation, the intersection of traditional models and innovative technologies is constantly giving rise to unprecedented business opportunities and industrial imagination space.

We have reason to believe that the "Doubao AI Phone" is not just a large model device with a new shell but transforms the phone into an intelligent entity that "can do things for you." Artificial intelligence has shown a glimmer of its commercial potential, indicating that a more intelligent and autonomous digital future is coming.

Is it a convenience or a threat that Doubao has learned to "take action"?

When AI steps out of the screen and holds our hands, security has become the most - concerned core topic.

As the AI assistant deeply embedded in the "Doubao Phone" system can perform functions such as automatic multi - platform price comparison, placing orders, and even obtaining game resources, the hidden dangers caused by its excessive permissions have begun to emerge.

Based on multiple reports, several Tencent applications (WeChat, Honor of Kings) and several Alibaba - affiliated applications (Taobao, Xianyu, Damai) have taken restrictive measures against it. For example, Honor of Kings has enabled AI control detection and only supports manual operation. Once the AI assistant intervenes, the account will be forced to log out. Taobao not only makes the automatic price comparison function unavailable, but even manual operation frequently triggers the security mechanism, resulting in login difficulties.

The series of reactions from all parties reflect the multiple challenges at the technical, commercial, and compliance levels that AI phones face during the implementation process. To cope with the increasingly tightening external environment, the Doubao Mobile Assistant issued statements on December 3rd, 5th, and 13th, emphasizing that its operations do not replace user authorization and do not intercept protected content, and calling for the establishment of clear usage specifications.

Meanwhile, the Doubao Mobile Assistant also issued a "Statement on Adjusting the AI's Ability to Operate the Phone," stating that it will restrict usage scenarios such as score - brushing and incentive - brushing, and further standardize operation permissions in financial and some game applications, aiming to seek a good fit among technological development, industry acceptance, and user experience.

Objectively summarizing the current challenges faced by the large - scale implementation of AI phones, they mainly focus on three aspects:

Security risks: Privacy and data storage become the core disputes

The unmanned operation of AI phones relies on the "perception - decision - execution" closed - loop, which requires the use of highly sensitive data including payment, identity, and operation records, and may transmit information back to the manufacturer's server. Although Doubao officially states that relevant data is not stored in the cloud and is not used for training, experts point out that once the server is attacked or there are internal management loopholes, users' control over their digital lives will be directly threatened.

Platform resistance: The battle for the entry point has begun

The AI phone uses the operating system as a unified entry point and directly completes operations such as price comparison and placing orders through intelligent agent scheduling, without users having to open specific apps. This essentially weakens the user reach ability and commercial ecological value of the original application platforms, demoting them from the "entry point" to the "supply pipeline." Therefore, several applications such as WeChat, Taobao, and Alipay quickly took restrictive measures against it to prevent their entry - point status from being undermined.

Compliance dilemma: Finding a balance between convenience and risk

Currently, how AI phones comply with the user agreements of various platforms in automated operations, how to obtain permissions through compliant interfaces, and how to perform tasks without touching security risk control have all become difficulties in actual implementation. Doubao's official active adjustment reflects that in the absence of clear rules for data boundaries and operation permissions in the existing ecosystem, enterprises have to actively shrink their front lines to gain survival and development space.

Overall, the restrictions faced by the "Doubao AI Phone" are not just technical adaptation issues but also touch on the deep - seated game of data sovereignty, platform interests, and ecological dominance. If AI phones are to be truly popularized, they need to find sustainable solutions under the protection of user privacy, platform cooperation mechanisms, and a clearer compliance framework. The industry's acceptance and the establishment of rules still depend on the continuous dialogue and joint exploration of technology providers, platform providers, and regulators.

A dialogue about execution, ecosystem, and rules

Through the heated discussions triggered by the Doubao AI Phone, we can see that it is not just about the success or failure of a single product but also a key signal for the entire AI industry to enter the deep - water area of implementation.

From technology demonstrations to practical tools, from industry topics to public attention, the pursuit, doubts, debates, and cooperation triggered by the Doubao AI Phone together outline the trends that AI must face after entering the deep - water area:

1. From "information assistant" to "execution partner"

In the past, AI mainly stayed at the level of dialogue and information processing, playing the role of an auxiliary tool. However, the emergence of the Doubao Phone has brought about a significant transformation. Its demonstrated capabilities enable AI to truly perform operations and complete tasks, such as booking tickets, comparing prices, and playing games, achieving a leap - forward breakthrough in AI auxiliary functions.

This transformation means that the future development direction of AI phones will undergo a qualitative change, shifting from simply competing in the number of functions to truly focusing on "whether AI can help me solve problems and achieve goals." People no longer need to perform cumbersome manual operations. They just need to tell AI their needs, and AI can complete complex operations on their behalf, allowing people to focus more on goals and decision - making.

In the future, when evaluating the quality of a product, it may no longer be just about the number of functions or the flashiness of the interface, but about whether it can be smoothly invoked by AI and accurately execute tasks. This will also force the software industry to redesign products and promote the formation of a new set of "AI - friendly" standards.

2. Platform competition shifts from "traffic competition" to "ecosystem integration"

Currently, the restrictions and blockades faced by the "Doubao AI Phone" seem to be functional conflicts on the surface, but actually behind them is the core game of "who will be the entry point and who will control the ecosystem." In the short term, the game between the two sides will continue.

But in the long run, users' needs will be gradually met in this collision. In the future, mobile phone manufacturers, AI companies, and service platforms are likely to move towards a model of "open cooperation and value sharing."

3. Technological innovation must go hand in hand with security and compliance

The privacy, data, and platform security issues exposed by the Doubao Phone actually serve as a wake - up call for the entire industry: the more powerful and deeply integrated AI is into our lives, the more severe the challenges of security and compliance will be.

In the future, the industry's development must find a balance between convenience and security and cannot be achieved at the cost of sacrificing user privacy or violating platform rules. This requires the joint efforts of regulators, enterprises, and technology providers to establish clear data usage specifications, permission boundaries, and responsibility mechanisms, and draw a runway for the healthy development of AI.

Overall, the success of the "Doubao AI Phone" is not the victory of a single enterprise but the result of the entire innovation ecosystem. From chip suppliers, algorithm developers, application creators to end - users, an open and collaborative value network has been formed. This ecosystem model is the ideal environment for the development of new - quality productivity.

How far is the "last mile" for large models to be implemented?

Who Moved My Cheese? is a fable novel written by Spencer Johnson. The book reveals through a fable that "the earlier you let go of the old cheese, the earlier you will find the new cheese."

Similarly, today when AI phones are moving from concept to reality, security and compliance are not an invisible high wall blocking the industry's development but a necessary path for it to mature and establish a new order.

The pursuit and controversy surrounding the "Doubao AI Phone" are precisely a microcosm of this process. It reflects the deep - seated transformation of large model technology from concept to practical use, from tool to partner. This discussion is essentially about how large model technology can be "implemented."

As witnesses and recorders of this transformation, we need to deeply consider whether AI phones can cross the "concept hype" and achieve large - scale popularization, which will determine their real market future.

Currently, the industry is at a critical turning point from novelty - seeking to practical use. Users are no longer satisfied with the "AI - enhanced" marketing claims but are more concerned about whether it can be integrated into real scenarios and solve practical problems.

AI phones should not stop at dialogue and response but should be able to understand intentions and provide active services, achieving an efficient end - cloud collaborative experience. At the same time, the cost should be controlled within a reasonable range to make technology accessible to the public. AI phones should not be information islands. Only by building an open, collaborative, and sustainable intelligent ecosystem can products achieve a technological leap from a single point to a broader scope.

This process is not only a technological breakthrough but also a systematic innovation in models, management, and systems. The business transformation driven by large models is not a zero - sum game but will give rise to new business forms and opportunities through continuous trial - and - error, cross - border integration, and problem - solving, promoting the efficiency leap of the entire business model.

Therefore, the future of AI phones does not lie in "whether there is AI" but in "whether AI is truly usable, easy to use, and useful" so that AI phones can go further in this marathon from concept to scale. Standing at the critical point of the intelligent revolution, the market's answer will ultimately be written in each user's actual choices.

This article is from the WeChat official account "Wandian Research" (ID: Agumanhua), author: Siyan, published by 36Kr with authorization.