SSPAI Morning Brief: Draft Rules Propose Timed Exit Reminders and Timely Intervention for AI Chat Addiction

Morning Brief

Draft rules propose requiring AI chat services to issue timed exit reminders and intervene promptly in cases of addiction
Xiaohongshu responds to a public “call-out” from the Lijiang Old Town Culture and Tourism Bureau
Cyberspace Administration issues regulations to standardize the behavior of online influencer accounts
Honor releases the WIN series gaming smartphones
Plaintiff presents forged evidence bearing Doubao AI watermarks, receives a judicial reprimand
Beijing arbitration case clarifies that dismissal due to a role being replaced by AI constitutes unlawful termination
Rumors You Can Just Glance At

Draft rules propose requiring AI chat services to issue timed exit reminders and intervene promptly in cases of addiction

On December 27, the Cyberspace Administration of China released the Interim Measures for the Management of Anthropomorphic Interactive AI Services (Draft for Public Comment), soliciting public feedback. The draft defines “anthropomorphic interactive services” as products or services that use AI technologies to simulate human personality traits, thinking patterns, and communication styles, and engage in emotional interaction with humans through text, images, audio, video, and other forms.

Beyond basic content compliance, the draft proposes that anthropomorphic interactive services must not provide false promises that seriously affect user behavior or services that harm social relationships; must not harm users’ physical or mental health; and must not induce users to make unreasonable decisions through algorithmic manipulation, misleading information, or emotional traps. Service providers are required to have safety capabilities such as mental health protection, emotional boundary guidance, and dependency risk warnings, and must not take replacing social interaction, controlling users’ psychology, or inducing addiction and dependence as design goals.

The draft also proposes introducing a number of intervention mechanisms. For example, while protecting users’ personal privacy, providers should assess users’ emotional states and their level of dependence on products and services, and take necessary measures to intervene when extreme emotions or addiction are identified. When users explicitly express intentions to commit suicide, self-harm, or other extreme actions, the conversation must be taken over by human operators, and timely measures should be taken to contact users’ guardians or emergency contacts (required to be provided during registration for minors and older adults). In addition, if a user has been using the service continuously for more than two hours, dynamic reminders—such as pop-up notifications—should be displayed to prompt a pause; providers must not obstruct users from actively exiting the service. Providers should also prominently inform users that they are interacting with AI rather than a natural person.

The draft further brings training data under regulatory oversight. Providers should strengthen the management of training data, ensure that data sources are lawful and traceable, take necessary measures to safeguard data security, and prevent data leakage risks. Unless otherwise stipulated by laws or administrative regulations, or with separate user consent, providers must not use user interaction data or users’ sensitive personal information for model training.

Since the beginning of this year, international media have repeatedly reported on incidents linked to AI chat services. The U.S. Federal Trade Commission is investigating chatbot products from major AI companies, and some states have begun to legislate oversight. For example, in October, the Governor of California signed a series of restrictive bills targeting AI applications, aimed at strengthening protections for minors. These measures require AI products to verify users’ ages, remind users every three hours that they are interacting with AI rather than a human, and promptly intervene in cases involving suicide or self-harm.

Xiaohongshu responds to a public “call-out” from the Lijiang Old Town Culture and Tourism Bureau

According to a report by Observer Network, the Culture and Tourism Bureau of Lijiang’s Old Town District in Yunnan recently issued a public letter accusing the Xiaohongshu platform of failing to adequately fulfill its responsibilities in information oversight. The bureau claimed this had led to a large number of consumers using unsubstantiated “avoid-at-all-costs” posts to threaten wedding photography businesses into issuing refunds.

In response, on December 24 Xiaohongshu customer service stated that the platform provides public reporting channels and will handle confirmed violations in a timely manner. People familiar with the platform said that preliminary reviews have not found evidence of consumers fabricating experiences or engaging in other violations. The Old Town District Culture and Tourism Bureau later deleted the letter. However, the Lijiang Old Town Travel Photography Industry Association said it has refused to continue negotiations with Xiaohongshu and is currently commissioning lawyers to collect evidence and prepare a collective public-interest lawsuit.

Several local travel photography business owners told the media that some consumers, after signing off on services and receiving their final photos, subsequently post negative reviews on Xiaohongshu to extort full refunds. In some cases, social groups have even emerged that coach members on how to obtain free services through complaints. Merchants said that even when they submit signed satisfaction agreements or certificates showing no complaints filed with regulators, their appeals often fail because the platform’s system deems_toggle “insufficient evidence.” As a result, order volumes at some studios have dropped sharply, and some sales teams have even resigned en masse. According to the industry association, as of December 2025, direct economic losses caused by false “avoid-at-all-costs” posts have exceeded RMB 1 million, with indirect order losses reaching more than RMB 5 million.

Public opinion online has been divided. Some users believe that certain “avoid-at-all-costs” posts have deteriorated in nature and have expressed concern about Xiaohongshu’s overall content environment. Another perspective shifts the focus to the travel photography businesses and the industry itself, arguing that the key to avoiding such posts lies in improving service quality.

Industry experts note that social platforms such as Xiaohongshu have effectively become part of the infrastructure for tourism services. Meanwhile, structural issues within the travel photography industry—such as cutthroat low-price competition, hidden fees, and non-standardized services—objectively increase the difficulty for platforms to verify the authenticity of content. Experts suggest that platforms should consider establishing collective appeal and communication mechanisms tailored to regional industries, while industry associations should help optimize the environment for public opinion oversight through standardization efforts.

Cyberspace Administration issues regulations to standardize the behavior of online influencer accounts

On December 26, the Cyberspace Administration of China (CAC) released the Notice on Regulating the Management of Online Influencer Account Conduct, establishing a negative list for online influencer account behavior and defining clear boundaries for such conduct.

According to the notice, the negative list of prohibited behaviors for online influencer accounts includes: (1) spreading vulgar content; (2) promoting unhealthy value orientations; (3) creating “ugliness-as-spectacle” or otherwise harmful personas; (4) disseminating false information; (5) distorting or misinterpreting policies and public events; (6) inciting group antagonism; (7) doxxing or publicly exposing individuals at will; (8) organizing offline fights or online flame wars; (9) soliciting negative leads or accusations; (10) calling on fans to gather or mobilize; (11) engaging in coercive or extortionary practices for personal gain; (12) conducting operations without required qualifications; and (13) covertly engaging in illegal or gray-market activities.

The notice requires all websites and platforms to conscientiously implement relevant laws, regulations, and policies, strictly fulfill their primary responsibilities in accordance with the negative list, and further improve community rules and user agreements related to the management of online influencer accounts. Platforms are also instructed to strengthen, in accordance with the law and platform rules, the management of content publishing, live-stream interactions, topic creation, and other behaviors by online influencer accounts, guiding operators to use their influence responsibly and to standardize online speech and conduct.

Honor releases the WIN series gaming smartphones

On December 26, Honor launched the WIN series smartphones in Chengdu, introducing two models—Honor WIN and Honor WIN RT—with starting prices of RMB 3,999 and RMB 2,699 respectively. The WIN series is a brand upgrade of the previous GT lineup and is positioned squarely in the professional gaming phone market.

In terms of performance, the entire Honor WIN lineup is powered by Qualcomm’s flagship Snapdragon platforms. The Honor WIN features the fifth-generation Snapdragon 8 Elite chipset, paired with LPDDR5X memory and UFS 4.1 storage. To improve endurance, the WIN series raises battery capacity to 10,000mAh for the first time, supporting 100W wired fast charging and 80W wireless fast charging (standard model only). The new phones also support bypass charging, allowing the charger to power the motherboard directly to reduce heat buildup while gaming and charging simultaneously.

The Honor WIN series introduces the “Dongfeng Turbo” active cooling system, which uses an ultra-compact fan and a direct-drive high-efficiency cooling air duct. Official figures claim a 30% improvement in cooling efficiency, with fan noise as low as 25 dB. On the display front, the phones are equipped with a 185Hz ultra-high-refresh-rate esports screen, support 5,920Hz high-frequency PWM dimming, and are optimized for ultra-frame-rate modes in multiple mainstream games.

The Honor WIN series also features dedicated esports antennas to ensure low-latency performance in weak network conditions. Additional features include AI-assisted control enhancements and anti-mis-touch functions. At the launch event, Honor announced a strategic partnership with Delta Force, with the WIN series designated as the official device for its professional league. Honor stated that it will further build out a gaming ecosystem around the WIN series, including future products such as tablets.

Plaintiff presents forged evidence bearing Doubao AI watermarks, receives a judicial reprimand

Recently, the Civil Division II of the Dawu County Court in Xiaogan, Hubei Province, identified a plaintiff’s attempt to fabricate evidence using AI technology during the trial of a housing lease contract dispute, and issued a formal reprimand in accordance with the law.

In May 2024, Li signed a housing lease contract with Xiong, agreeing on the lease term and rent, and specifying that the rent for the latter half of the term must be paid by March 2025. After the lease expired, Li sued Xiong for failing to pay the remaining rent as well as water and electricity fees, appointing his daughter Dong as his litigation agent.

During the hearing, Li was unable to provide valid proof that Xiong owed water and electricity charges. Dong stated in court that Xiong was the first tenant of the property and that all utility costs were incurred by him, and promised to submit photos of the water and electricity meters from before and after the lease period as supporting evidence after the hearing. However, the photos Dong submitted clearly bore a “Generated by Doubao AI” watermark, which immediately raised the judge’s suspicions. Further examination revealed inconsistencies in Dong’s statements: she initially claimed that the meters were used independently, but later admitted under questioning that two households shared the same meters. Faced with the judge’s inquiries, Dong was unable to justify her claims and ultimately admitted to fabricating the evidence, subsequently submitting the genuine photos to the court.

The presiding judge gave Dong a serious lesson in the rule of law, pointing out that using AI-generated images as litigation evidence constitutes the fabrication of key evidence and is suspected of disrupting judicial order. Taking into account that Dong voluntarily confessed during the investigation and promptly corrected her mistake, the court followed a principle of combining punishment with education, opting to handle the matter by way of a reprimand and refusing to admit the fabricated evidence.

Internationally, the use of AI-generated false materials in litigation has also become increasingly common. A database established by French lawyer Damien Charlotin has collected more than 600 cases worldwide. When such errors are discovered, many lawyers attempt to deflect responsibility with various excuses, such as claiming they did not know they were using AI, did not realize AI could hallucinate, alleging that hackers remotely altered the files, or even saying it was too inconvenient to switch windows on a laptop so they used AI. Some lawyers have even tried to shift the blame to interns or clients.

Beijing arbitration case clarifies that dismissal due to a role being replaced by AI constitutes unlawful termination

On December 26, the Beijing Municipal Human Resources and Social Security Bureau released its 2025 annual collection of representative labor and personnel dispute arbitration cases. Among them, an arbitration case arising from AI replacing a job position explicitly clarified that “AI replacing a position does not equal lawful dismissal.”

According to the case details, Liu had worked for many years at a technology company, responsible for traditional manual map data collection. In early 2024, the company decided to fully shift to AI-driven automated data collection, abolishing the department and corresponding positions where Liu worked. At the end of 2024, the company terminated Liu’s labor contract on the grounds that “the objective circumstances on which the labor contract was based have undergone a material change, making the contract impossible to continue performing.” Liu believed the termination was unlawful and therefore applied for arbitration.

In its analysis, the arbitration committee pointed out that the company’s introduction of AI technology was a proactive technological upgrade undertaken to adapt to market competition. Although such changes may lead to adjustments in job structures, they do not meet the criteria of “objective circumstances,” which require characteristics of irresistibility and unpredictability. Terminating a labor contract on the grounds that a position has been replaced by AI essentially shifts the normal risks of technological iteration onto the employee. The arbitration committee thus determined that the company’s action constituted unlawful termination.

The arbitration committee further noted that, in the context of rapid technological development, employers should give priority to properly accommodating affected employees through measures such as negotiating changes to labor contracts, providing skills training, and arranging internal job transfers. If termination of a labor contract is indeed necessary, employers must strictly comply with relevant regulations and avoid mechanically applying “material changes in objective circumstances” as a justification for dismissal.

Rumors You Can Just Glance At

According to the November 2025 Domestic Mobile Phone Market Operation Analysis Report released by the China Academy of Information and Communications Technology (CAICT), domestic brands shipped 23.228 million units in November, down 12.6% year over year, accounting for 77.7% of total shipments. Meanwhile, shipments of foreign-brand phones reached 6.93 million units, up 128.4% year over year. This surge is likely related to adjustments by manufacturers such as Apple in promotions, channel strategies, or new product release cycles.
Weibo user “定焦数码” claimed that “a new iteration of the iPhone Air is confirmed to launch and will appear at the fall event. In addition, the iPhone 17e has already entered mass production and will be seen at the spring event.” This differs from earlier reporting by The Information, which stated that due to lukewarm market performance of the first-generation iPhone Air, Apple had informed its supply chain to remove the second-generation product from the 2026 roadmap, possibly delaying it until 2027.
X user Kosutami recently shared a set of photos said to be prototypes of the first-generation AirPods, showing two bright color options Apple once experimented with: pink and yellow. The prototypes were described as using an “iPhone 5C color scheme.” Although these colorways never made it into mass production, the level of finish on the prototypes suggests that Apple seriously considered breaking away from its single-white design tradition during the early development stage.

SP-Life