Challenges in the Development of the AI Economy

1. Technical “Necklace” Risks

With the extensive application of AI technology, the demand for large models has surged. Training and running these models require massive computing resources and data storage capabilities, leading to an increasingly severe phenomenon of technical “necklace”.

On the one hand, the cost of high-performance computing equipment remains high, while on the other hand, data security and quality control have become key issues.

For example, the training costs of some large language models like OpenAI’s GPT-4 and Google’s Gemini Ultra reached $78 million and$191 million, respectively, making resources increasingly concentrated in the hands of a few giants. Although open-source models have lowered the entry barrier, they still lag behind closed-source models in performance. In the future, with the advancement of technology, we expect to see more cost-effective solutions and continued improvement in the performance of open-source models.

2.Difficulties in Technology Implementation

Although AI technology possesses strong potential in theory, it faces many challenges in practical application. Many projects perform well in the laboratory stage but fail to achieve the expected results when put into practical use due to low matching between technology and real-world scenarios.

For instance, some autonomous vehicles perform poorly under specific weather conditions, or some medical diagnostic systems may misdiagnose due to data bias. These problems highlight the difficulty of transforming from the laboratory to the market.

Palantir, a company known for data analysis and software development, has widely used its technology in the military and intelligence fields. However, when trying to commercialize these technologies and apply them to other industries, Palantir faced some difficulties. One reason is that its technical solutions usually require high customization, which increases costs and prolongs deployment cycles. Additionally, customers in non-military fields may not have the same security requirements and budgets, leading to challenges in technology implementation.

1.International Technology Blockades and Counter-Globalization Trends

近年来，以美国为首的发达国家掀起了一股技术封锁和逆全球化的浪潮，这对全球AI产业造成了严重影响。这种封锁不仅体现在对关键技术的出口管制上，还包括对国际科技合作的限制。

2024年3月，美国升级芯片出口禁令，表示将对中国出口的AI半导体产品采取“逐案审查”政策规则，全面限制英伟达、AMD等公司的先进AI芯片和半导体设备售往中国，同时计划限制中国AI大模型厂商通过美国云服务厂商使用海外算力，导致中国企业在获取高端芯片、技术、算力方面受到了限制。面对这一挑战，中国正采取多种策略来降低对外部技术的依赖，包括加大自主研发力度、寻求与其他国家的合作等。

2.Public Issues and Ethical Problems

The development of AI technology has triggered numerous public issues and ethical problems. On the one hand, personal privacy protection has become a focus of attention, especially in the process of data collection and use. On the other hand, the application of AI technology may have a profound impact on the job market, with automation potentially leading to the disappearance of certain positions. Additionally, issues such as algorithmic bias, lack of transparency, and unclear responsibility attribution are becoming increasingly prominent.

On July 7, 2024, an accident occurred involving a self-driving taxi operated by Baidu’s autonomous driving travel service platform “Luobo Kuaipao” in Wuhan, where the vehicle collided with a pedestrian. The complex issues surrounding responsibility determination and compensation highlight the importance of formulating clear laws and regulations to regulate the application of AI technology.

1.Market Fluctuations in Capital Markets

Despite attracting substantial investment, the capital market for AI remains highly volatile. Many startups, after experiencing initial rapid growth, struggle due to a lack of profitability or unclear business models, leading to layoffs or closures.

Pure AI algorithm companies are particularly susceptible to market fluctuations, heavily influenced by investment sentiment. For instance, in early February, due to market speculation surrounding the AIGC concept, the stock prices of three algorithm companies surged. However, they subsequently declined due to uncertainty about long-term performance growth. Additionally, some companies experienced abnormal stock price fluctuations due to false advertising or exaggerated efficacy claims, reflecting the high risk and uncertainty of the AI industry and exposing the current immaturity of AI market business models.

2. Fluctuating Market Demand and Scale

The market demand for AI technology is characterized by volatility. On the one hand, as technology matures, more and more enterprises begin to adopt AI solutions to improve efficiency and competitiveness. On the other hand, market acceptance of AI products is not stable, with some consumers and businesses adopting a cautious attitude towards new technologies.

Again, referring to the aforementioned case of the autonomous driving accident, despite the significant progress made in autonomous driving technology, public concerns about safety remain a significant barrier to its promotion. Fluctuations in market scale and prices pose considerable operational pressures on enterprises.

1.Technological Innovation and Practical

Technological Innovation and Practical Implementation To address the aforementioned technical challenges, increased research and development investment is necessary to drive technological innovation. This includes improving computing efficiency and reducing energy consumption, as well as enhancing the generalization and robustness of models. Additionally, strengthening infrastructure construction, such as optimizing data centers and cloud computing platforms, is crucial for promoting the popularization and application of AI technology. Currently, the Chinese government has launched several programs to support the construction and upgrading of AI infrastructure, promoting the commercialization process of AI technology.

To effectively solve the difficulties in AI technology implementation, especially in improving the matching of technology with real-world scenarios, it is essential to focus on specific industry needs. Providing customized AI solutions for different industries is key. In May 2024, Gao Nianshu, Chairman of AsiaInfo Technologies, stated that “the gap between the implementation of general large models in vertical industries, the lack of unified management leading to repeated construction, and the lack of methods and tools for rapid implementation” are the main obstacles to the widespread application of large models. AsiaInfo Technologies summarized industry experience, based on real scenarios, and developed a vertical AI product system - the Yuan Si large model, providing customized AI solutions to address practical enterprise problems.

Technology providers need to continuously optimize and improve AI technology, enhancing its stability and accuracy. Simultaneously, they should focus on collaborating with enterprises to understand their needs, thus guiding the direction of technological development. Applying AI technology to specific scenarios can better identify promising patterns and models.

2.Talent Cultivation and Educational Reform

The development of AI technology is inseparable from a high-quality talent pool. Currently, Chinese enterprises still need to improve their maturity in AI strategy, especially lacking sufficient key talents like AI translators. Therefore, the education system needs to undergo reforms to increase AI-related courses and provide more practical opportunities, cultivating professionals to meet market demands. Additionally, strengthening interdisciplinary cooperation and integrating AI knowledge with other fields can help improve the overall social innovation capability. Currently, the Ministry of Education is actively promoting the popularization of typical application scenarios of “AI + higher education,” utilizing intelligent technology to support the innovation of talent training models and teaching methods.

3.Policy Support and International Cooperation

Policy Support and International Cooperation The government should introduce more supportive policies to encourage enterprises to participate in international technological cooperation and collectively address the challenges brought by AI technology. This includes simplifying regulatory processes and providing financial subsidies. At the same time, international cooperation is crucial for promoting the healthy development of AI technology.

In July 2024, the 78th United Nations General Assembly passed a resolution proposed by China on strengthening international cooperation in AI capacity building, marking a significant step forward in the international community’s efforts to promote cooperation and development in the AI field. This international cooperation not only promotes exchanges and sharing of AI technology among countries but also provides new ideas and frameworks for global AI governance.

Conclusion

about us