Analysis of the Impact of China's New Regulations on AI Anthropomorphic Interaction Supervision on Industry Investment

On December 27, the Cyberspace Administration of China issued the Interim Measures for the Management of Artificial Intelligence Anthropomorphic Interaction Services (Draft for Comments), which explicitly incorporates “anthropomorphic interaction” AI services into the regulatory system. It requires service providers to bear safety responsibilities throughout the life cycle, establish mechanisms for algorithm review, data security, and personal information protection, and prohibit behaviors such as endangering national security, false promises, emotional manipulation, and inducing user decisions [1]. The draft also requires active prompts to users to avoid addiction during operation, the establishment of intervention mechanisms, and the provision of human customer service, reflecting a high degree of caution towards the “personification” tendency of emotional interaction AI.

Anthropomorphic interaction is an important selling point for current differentiated products such as digital humans, virtual anchors, and human-machine co-creation. The new regulations require the establishment of multi-dimensional compliance capabilities such as algorithm evaluation, identity authentication, and human transfer, which will directly increase R&D and operation costs, especially putting pressure on small and medium-sized manufacturers. Investors need to re-evaluate the short-term cash flow prospects of “persona innovation”-driven companies (such as digital human output platforms), considering the dilution of gross profit margins and free cash flow by compliance investments.

The new regulations emphasize that service providers must bear the full-chain responsibility, which will become the industry’s access threshold and reverse expel lightweight products lacking risk control systems. Industry leaders with mature safety governance and data governance capabilities (such as iFLYTEK, SenseTime, etc.) are expected to form barriers in the “compliance and trustworthy” aspect, and their long-term value may be re-evaluated. Taking current market valuations as an example, iFLYTEK has a market value of about 112.9 billion yuan, showing stable capital market recognition; although SenseTime still has short-term profit pressure, its leading position helps attract stable funds [0].

Anthropomorphic AI has high stickiness in scenarios such as marketing, content production, and customer service. The new regulations restrict practices such as emotional manipulation and inducing decisions, forcing products to transform from “emotional connection” to “functional efficiency”. Investors should pay attention to whether enterprises refocusing on tool attributes can maintain user activity and payment rates, especially whether their commercialization paths rely on “personification” packaging. Some segment players with virtual idols and chat companions as core monetization methods face the risk of repositioning.

Although short-term compliance costs rise, the clear regulatory framework provides predictability for the industry instead. Compliance capabilities become a new competitive advantage, helping to enhance external capital confidence and cross-industry expansion. For example, scenarios such as medical care, educational assistance, and government consulting all emphasize safety and transparency more. Investors can pay attention to companies with a “trustworthy AI” base and actively expanding vertical industry applications, and predict their long-term market share growth in segmented markets.

• Carefully screen high-emotion-dependent projects
  : Focus on whether the business model over-reliance on emotional interaction and immersive experience. If the innovation cannot be depersonalized, growth may be limited after future regulatory restrictions.
• Focus on tracking compliance investment and governance output
  : Evaluate capabilities such as algorithm review, content control, and manual response on a project basis, and quantify their contribution to customer satisfaction and brand risk.
• Allocate targets with “strong compliance capabilities + multi-scenario implementation”
  : For example, enterprises with large model capabilities, industry models, content review, and data governance experience are more likely to be regarded as “policy-friendly” assets by capital.

The new regulations are expected to increase the compliance threshold and operating costs of the entire anthropomorphic AI segment in the short term, putting profit pressure on players relying on “personification experience”; at the same time, its clear regulatory framework is also expected to strengthen the barriers of leading enterprises with safety governance capabilities and provide support for long-term value. Investors should balance short-term compliance risks and long-term competitive advantages in their portfolios, focus on targets with mature capabilities in governance, data security, and industry implementation, and be alert to business models that are difficult to break away from “emotion-driven” and rely on irrational interactions.

[0] Jinling API Data (Real-time quotes: iFLYTEK, SenseTime, Alibaba)
[1] Reuters - “China issues drafts rules to regulate AI with human-like interaction” (https://www.reuters.com/world/asia-pacific/china-issues-drafts-rules-regulate-ai-with-human-like-interaction-2025-12-27/)