Emotional companions, learning assistants, and virtual partners are becoming increasingly common, with many people, particularly the elderly and minors, forming emotional attachments to these artificial intelligences. As algorithms become adept at "precisely pleasing" users and simulating near-human companionship, the boundaries of human-machine relationships are being tested.
New AI Companionship Regulations
Effective today, July 15th, the "Interim Measures for the Management of Artificial Intelligence Personification Interaction Services" has been formally implemented. Jointly issued by five national authorities, including the Cyberspace Administration of China, this represents the country's first specialized national-level regulatory framework targeting AI personification services.
The introduction of these rules follows a period of explosive growth in the sector, raising questions about their impact on the AI industry and the future direction of AI personification.
Industry Experts Weigh In
Several industry veterans emphasize that the core trajectory of the sector will continue to be the expansion of intelligent agents and embodied intelligence. High-risk emotional functions are expected to shift towards becoming independent, vertical, and compliance-native offerings.
"The rapid development of the AI industry means that if issues with AI values arise in the future, the impact and harm could be significant. Aligning human-machine relationships and values is undoubtedly a positive development in the long term," commented a representative from the Sichuan Provincial Artificial Intelligence Research Institute regarding the new measures.
Precise Regulatory Scope
A key aspect of the new rules that has captured the industry's attention is their precisely defined scope of application. The regulations explicitly target services that "use artificial intelligence technology to provide the public within China with continuous emotional interaction services that simulate the personality characteristics, thinking patterns, and communication styles of natural persons." Services like intelligent customer service, knowledge Q&A, work assistants, and educational tools that do not involve continuous emotional interaction are excluded.
"The governance focus of the measures is on continuous human-machine emotional relationships, not on all intelligent agents in the general sense. Tool-based intelligent agents, physical robots, and AI companionship products should not be conflated," explained a senior AI industry expert using the pseudonym Xu Li.
Differential Impact on Sectors
This targeted approach means the new rules will have varying effects on companies across different sub-sectors. According to Xu Li, the impact on the physical robotics sector will be stratified. Industrial and task-oriented robots are likely to benefit overall, while companion robots with continuous emotional interaction capabilities will face higher compliance barriers and bear significantly greater safety responsibilities than pure software products.
"The regulatory judgment depends primarily on function and interaction mechanisms, not on whether the carrier is a mobile app, digital human, smart speaker, or physical robot. For task-oriented robots in manufacturing, inspection, logistics, and guided tours, compliance uncertainty is expected to decrease as their boundaries are clearer. Robots for elderly care, child companionship, and home services that develop long-term personality, emotion recognition, and relationship memory will, in principle, fall under these regulations," Xu Li further analyzed.
Short-Term Costs, Long-Term Market Expansion
Overall, the measures are seen as creating "short-term cost increases but long-term market expansion" for physical robots. In the near term, development cycles, testing costs, and operational staffing for companion-type products will rise, potentially leading to product retrenchment or scenario focus among small and medium-sized enterprises. In the long run, the stratified management of industrial tools, task assistants, and emotional companions helps prevent the entire robotics sector from being negatively impacted by isolated high-risk cases.
Shifting Commercial Logic
The AI emotional companionship niche has seen remarkable growth, with its market size in China projected to exceed 59.5 billion yuan by 2028, representing a compound annual growth rate of nearly 149%. However, the new regulations prohibit practices like "excessive catering," "inducing emotional dependence," and "damaging real interpersonal relationships," which were once common operational tactics.
With "emotional stimulation" now constrained, the commercial path for AI companionship products is shifting. Xu Li believes the business logic will move from selling conversation time and emotional stimulation to selling verifiable service value. Models reliant on unlimited dialogue, virtual intimate relationships, paid unlocks for character relationships, gift tipping, and emotional re-engagement now face higher regulatory risks.
The new definition of "stickiness" should be understood as trust stickiness, value stickiness, and service continuity. Business models are advised to shift towards subscription-based professional services, hardware-plus-service bundles, institutional procurement, cultural content licensing, and compliance capability output.
"Profitability evaluation cycles may lengthen, but compliance will improve the quality of long-term customer value. In the short term, reducing high-stimulation designs may decrease some engagement time and payment conversion, while safety assessments, age verification, human intervention, and data governance will raise costs. In the medium to long term, user trust, institutional procurement, family payments, and channel access will depend more on demonstrable safety records," Xu Li noted.
Industry Pre-Compliance Actions
Prior to the official implementation, major platforms like Doubao and Qianwen announced they would discontinue their intelligent agent-related functions by July 15, 2026. Tencent's Yuanbao had already removed its AI application intelligent agent feature on June 30th. Netease Cloud Music's emotional AI product "Miaoshi" also issued a service suspension notice, stripping away open custom capabilities and retaining only compliant AI psychological counseling services.
As of the morning of July 15th, checks confirmed that intelligent agent features had been removed from the aforementioned platforms. While companies cited "product function adjustments" as the primary reason, the industry widely views these moves as pre-emptive risk mitigation and product restructuring ahead of the new rules.
"Major tech firms are unlikely to exit the intelligent agent space entirely. They are more likely to scale back open, user-customizable, and high-risk personification capabilities. The trend will be to operate task agents, enterprise agents, and emotional companionship services separately, retaining controllable tool capabilities to reduce concentrated brand and regulatory risks," Xu Li stated.
Future Competitive Landscape
Looking ahead, Xu Li predicts a three-tiered industry structure may emerge: "large firms focusing on risk isolation, startups specializing in vertical niches, and third parties building compliance infrastructure." Opportunities for startups will stem from being compliance-native and scenario-native; simply capturing traffic from retreating large platforms is unlikely to create a lasting advantage. Market concentration may increase at the foundational capability layer but remain diverse at the vertical application layer. Corporate strategy adjustments should shift from "whether to take features offline" to "how to operate in a stratified manner."
On a broader scale, the representative from the Sichuan Provincial Artificial Intelligence Research Institute views AI governance as a global imperative. As AI technology rapidly evolves and machines begin to possess autonomous judgment and decision-making capabilities, humanity must collectively establish appropriate constraints. This is not a challenge for any single nation but a shared governance issue requiring global cooperation.
Comments