24-hour AI news at a glance: Domestic regulation and education policies are advancing simultaneously, and overseas focus has shifted to cutting-edge model risk assessment and AI computing power expansion.
In the past 24 hours (April 12, 2026 to April 13, 2026), AI news has presented "two main lines" in parallel: domestic emphasis is placed on governance rules, education system construction and industry implementation, while foreign countries focus on cutting-edge model security, government cooperation and chip production capacity expectations. On the whole, AI competition has further extended from simply building models to systems, infrastructure, and vertical scenario applications.
1. The five departments issued the management measures for anthropomorphic interactive services
The Cyberspace Administration of China and other five departments jointly issued the "Interim Measures for the Management of Artificial Intelligence Anthropomorphic Interactive Services", clarifying that relevant services shall not provide high-risk content such as virtual intimate relationships to minors, and requiring strengthening risk warnings. This move shows that supervision has further penetrated from generative content management to the level of "emotional companionship" and "personality simulation" AI applications.
2. The education system has entered the stage of refinement of "artificial intelligence + education"
Focusing on the "Artificial Intelligence + Education" Action Plan, the latest interpretation of the education authorities emphasizes the integration of AI into talent training, curriculum system, teacher ability and educational governance. For the industry, this means that AI education is no longer just a pilot project, but a transition to a larger range of institutionalization.
3. 16 Chinese science and technology societies jointly released the Global AI Governance Initiative
On April 13, 16 scientific and technological societies, including the Chinese Society of Automation and the Chinese Computer Society, jointly issued an initiative proposing people-oriented, intelligent goodness, and opposing technological hegemony and exclusive "small circles". This reflects that domestic academia and industry are taking the initiative to strengthen international governance discourse and cooperation frameworks.
4. The hospital's exclusive "AI operating system" has been applied in real scenarios
Southern Hospital of Southern Medical University and Huawei released the hospital general artificial intelligence platform, and disclosed that the system has covered nephrology, anesthesia, pathology, medical records and other links. Medical AI is moving from a single point of assistance to platformization and full-process collaboration, which is of more practical significance to the digital upgrading of hospitals.
5. Hong Kong has proposed a significant increase in computing power to compete for the position of a global AI hub
Hong Kong said it will take the significant increase in supercomputing capabilities as the starting point to promote itself to enter the forefront of global AI hubs. Computing power, capital and policy synergy are becoming the core of regional competition, and the next key for Hong Kong is whether it can turn its infrastructure advantages into industrial agglomeration effects.
6. The UK financial regulator urgently assesses the risks of Anthropic's new model
The Bank of England, the Financial Conduct Authority and other institutions are assessing the potential cybersecurity and infrastructure risks posed by Anthropic's latest model. The capability boundaries of cutting-edge models are getting closer and closer to the safety red line of critical systems, and regulatory authorities have begun to intervene in advance, indicating that the path of "deploy first and evaluate later" is being rewritten.
7. Anthropic signed a memorandum of understanding on AI security cooperation with the Australian government
Anthropic announced a cooperation with the Australian government to promote AI safety research, national AI program goals, and scientific research cooperation, and support medical diagnostics and educational research. Cooperation between AI companies and sovereign governments is escalating from general policy exchanges to more specific joint projects.
8. TSMC is driven by AI demand, and the market expects profits to reach a new high in the first quarter
The market generally expects TSMC to achieve record profitability for the fourth consecutive quarter, driven by strong demand for AI chips. The chip link is still one of the most definite beneficiaries of the AI industry chain, which also shows that the computing power arms race has not yet shown obvious signs of cooling.
9. OpenAI updates enterprise functions, and ChatGPT can be written directly into Microsoft and Google office applications
OpenAI's latest Enterprise Edition update shows that ChatGPT has supported writing operations to some Microsoft and Google Office applications, such as drafting emails, creating documents and forms, and scheduling meetings. AI assistants are moving from "giving suggestions" to "directly executing tasks", and the integration speed of enterprise workflows has accelerated significantly.
Frequently Asked Questions (Q&A)
Q: What is the most obvious trend in the AI industry in the past 24 hours?
A: The most obvious trend is that "model competition" is expanding into a comprehensive competition of "supervision, computing power and scenario implementation".
Q: Which side of domestic news is more focused?
A: China is more inclined to governance rules, education system and medical applications, emphasizing controllable implementation and long-term infrastructure.
Q: Which side is the focus of foreign news more biased?
A: Foreign countries pay more attention to the security of cutting-edge models, government cooperation mechanisms and high-end chip supply, reflecting the governance pressure after capability spillover.
Q: What does this mean for developers and entrepreneurs?
A: The opportunity is still there, but the direction is shifting from simply doing models to making compliant products, vertical applications, and tools that are deeply integrated with existing workflows.
