Xinhua News Agency, Beijing (Reporter Sun Jing and Hu Dandan) With the rapid development of artificial intelligence (AI) technology, multiple manufacturers are attempting to launch artificial intelligence personal computers (AI PCs). Industry experts believe that this combination will bring unprecedented intelligent experiences and productivity improvements, but it also faces a series of technological and market challenges. It is worth observing whether the combination of AI and PC can accelerate the entry of AI into ordinary households.

Multiple PC giants enter the market to lay out AI PCs

Microsoft released Copilot+PC specifically designed for AI in May this year. Copilot is a cross platform AI assistant launched by Microsoft, which is a key part of Microsoft's promotion of artificial intelligence strategy. Copilot+PC is known as an innovative product aimed at redefining the way personal computers interact with users. Currently, well-known PC manufacturers such as Lenovo, Dell, HP, and Samsung have signed cooperation agreements with Microsoft.

Google subsequently announced the addition of new AI features to its high-end Android laptop Chromebook Plus product line. In March, Apple launched the MacBook Air series equipped with the M3 chip, emphasizing the artificial intelligence attributes of the product.

In China, Lenovo Group released an AI PC series product with a built-in personal intelligence agent called "Lenovo Xiaotian" in April. Companies such as Xiaomi and Huawei are also exploring or have launched personal computers with AI features.

According to Gao Yu, General Manager of the Technology Department of Intel Corporation in China, the six major application scenarios of AI PC are: chat robots bringing a more professional Q&A experience, local knowledge bases allowing knowledge wealth to be shared and utilized, personal assistants helping users efficiently complete daily affairs, office assistants efficiently processing text and even drafting contract documents, multimedia assistants helping to process video and image content, and PC management making remote management and protection more efficient. These applications can help businesses reduce costs and increase efficiency.

Is AI PC a true revolution

At present, major PC manufacturers, chip manufacturers, and operating system providers have different definitions of AI PCs, but they all have clear requirements for computers to efficiently process AI tasks.

According to Microsoft's standards, AI PCs need to be equipped with hardware that can provide powerful neural network processing capabilities, such as neural network processing units (NPUs) designed specifically for AI computing tasks, to meet the needs of large-scale model operation. At the same time, computers require sufficient random access memory. In addition to the aforementioned hardware requirements, AI PCs also need to have good system optimization and software support.

Gao Yu stated that AI PC is the "third revolution" of PC, bringing revolutionary changes to user experience and productivity through AI.

Pan Chunchun, Chairman and CEO of Zhitong International Information Technology Co., Ltd., said in an interview with Xinhua News Agency, "AI PC represents a new type of personal computing platform, a brand new product transformation; it is the ecological carrier of 'intelligent agent assistant+PC hardware+full terminal' that transforms the core product attributes from a single carrier of 'application software+basic software+PC hardware'."

Dual blessing of "computing power+privacy protection"

AI systems need to process a large amount of user data, which may involve sensitive information such as personal privacy and trade secrets. Pan Chunchun said when discussing the difference between AI PCs and traditional PCs, apart from computing power, for many users, "the local AI computing power of AI PCs should be a significant difference between AI PCs and traditional PCs.".

"Users can still smoothly run AI tasks in offline environments, ensuring data privacy and security. This local offline AI computing capability is one of the significant advantages of AI PCs over 'cloud based AI+traditional PCs'," he said.

Liu Huifu, Chief Technology Officer of Softcom Power Group, believes that in addition to being systematic, scenario oriented, and vertical, "localization" is the core feature of AI PC. AI PC emphasizes the ability of local end side models, which can better achieve real-time interaction, privacy protection, and hardware full connectivity of AI.

Some technology bloggers believe that large models must be directly integrated into consumer electronics devices such as smartphones and computers in order to truly achieve personalized applications. Due to privacy concerns, many users are hesitant to upload their data to the cloud. The characteristics of AI PC enable it to directly run end-to-end models locally, and the models themselves can be deeply integrated with computer systems, making data usage safer and faster, which is also the advantage of AI PC.

Security Challenges and Popularization

The launch of AI PCs has not been smooth sailing, and the first and foremost issue is the inherent security issues of AI.

Microsoft recently announced that it will temporarily postpone the launch of an AI feature called Recall designed for Copilot+PC. The Recall function allows users to quickly retrieve information they have browsed in applications, websites, images, and documents, building a searchable visual timeline. Many people are concerned that once hackers exploit vulnerabilities to invade, the user's computer activity records will be exposed.

Xue Zhihui, an expert of the Artificial Intelligence Security Governance Committee of the China Cyberspace Security Association and Vice President of Tianrongxin Technology Group, pointed out in an interview with Xinhua News Agency that AI PCs may also face risks such as bias or errors in AI decision-making, unfair or adverse effects on users. For example, imbalanced training corpora from different countries may lead to conflicting values, so the decision-making solutions provided by artificial intelligence may not necessarily be applicable to all users.

Another issue that needs to be addressed in promoting AI PC is the real iterative needs of users. A survey released by Reuters at the end of May showed that the usage rate of generative artificial intelligence tools is still low. Researchers conducted a survey of over 10000 people in six countries, including the UK, the US, Argentina, Denmark, France, and Japan, and found that the usage rate of ChatGPT, the most widely recognized generative artificial intelligence tool, is still very low. Only 1% of people in Japan use it every day, 2% in France and the UK, and 7% in the US.

In addition, due to the combination of new chips and software, newly launched AI PCs generally have higher costs and high prices. Although the combination of AI and PC has a promising future, it still takes some time to enter the homes of ordinary people. (Participating journalists: Dong Yalei, Wu Huijun, Sun Lei)