New Standards for AI Terminals

On May 8, the series of national standards titled “Intelligent Classification of AI Terminals” (GB/Z 177—2026) was officially released.

AI terminals are key carriers for the large-scale implementation and systematic development of AI technology. In recent years, China’s AI industry has flourished, with AI terminals driving new products, business models, and experiences, effectively stimulating consumer enthusiasm and becoming an important lever for boosting domestic demand and optimizing consumption structure.

Upgrading from Passive Tools to Intelligent Assistants

Wei Ran, Chief Engineer of the China Academy of Information and Communications Technology, explained that AI terminals are a new generation of intelligent terminals driven by large models. Compared to traditional terminals, AI terminals can actively perceive scenarios, accurately understand user intentions, and possess multimodal interaction capabilities, including text, voice, and audio-visual.

Additionally, AI terminals support generative applications and intelligent agent services based on personal large models and knowledge bases, enabling them to learn autonomously and evolve continuously.

“Overall, intelligent terminals are transitioning from traditional passive execution tools to perceptive, understanding, service-oriented, and growth-capable intelligent assistants, redefining the human-computer interaction relationship. These capabilities are also the core functions emphasized in the highest level of the new intelligent classification national standards,” said Wei Ran.

New Standards Overview

Regarding the new standards, Yu Xiuming, Deputy Director of the China Electronics Standardization Institute, introduced that the series adopts a “2+N” framework. The “2” refers to “Part 1: Reference Framework” and “Part 2: General Requirements,” which clarify the concept of intelligence, classification levels, and testing methods, addressing fundamental questions such as “What is an AI terminal? How is it classified? How is it evaluated?” This forms the foundation for all category standards.

The classification system ranges from L1 response level, L2 tool level, L3 assistant level, to L4 collaborative level, with the intelligence level continuously enhancing, making the terminals increasingly “smart.” The “N” refers to specific standards for different products such as smartphones, computers, TVs, smart glasses, automotive cabins, speakers, and headphones.

Three Development Paths

Yu Xiuming noted that research and testing analysis show that products with high user ownership rates are generally at L1 and L2 levels, while some new products can reach L3 level.

“Currently, AI terminal forms are flourishing, evolving along three parallel paths: upgrading traditional terminals, expanding emerging terminals, and exploring future terminals,” Wei Ran stated.

On one hand, traditional general-purpose terminals are first upgraded to AI terminals, with shipments of AI smartphones, PCs, and tablets surpassing ten million units, becoming the current market mainstay. On the other hand, emerging categories such as AI in-vehicle terminals, smart glasses, and AI toys are rapidly growing.

Simultaneously, AI-native terminal forms represented by embodied intelligence are also continuously being explored, further unleashing the potential of AI applications.

Future Directions

“Currently, AI and terminal technologies are undergoing systematic integration, and future breakthroughs need to focus on three major directions,” Wei Ran indicated: 1) optimizing the end-cloud collaborative architecture, where the cloud handles high-complexity tasks while the end processes high-frequency real-time interaction tasks for efficient collaboration of intelligent computing; 2) deepening full-stack upgrades of hardware and software, enhancing core capabilities such as computing, storage, and perception on the hardware side, while promoting AI capabilities from the application layer down to the operating system layer, building a system-level AI foundation to better support innovative applications of intelligent agents; 3) upgrading security and privacy protection systems to solidify data security and privacy barriers on the end side, ensuring that AI terminal services are trustworthy, secure, and controllable throughout the process.

Yu Xiuming stated that the establishment of the “Intelligent Classification of AI Terminals” series of standards will provide enterprises with clear benchmarking paths and improvement directions, helping to promote high-end product supply, enhance resource utilization efficiency, and foster orderly competition and healthy development in the supply side.

At the same time, the standards will also provide consumers with clear technical references and evaluation criteria, allowing the demand side to have clear benchmarks to make more rational choices for their intelligent products, further enhancing user experience and satisfaction.

It is reported that relevant departments will continue to advance standard development in areas such as wearable devices, home appliances, and trendy toys, promoting the gradual realization of comprehensive coverage of intelligent classification for various terminals, allowing more products to have clear standards.