Changxin Fund's Shen Wei: Alpha Returns May Decline in the AI Era; AI Effectiveness Depends on Data Quality

In June this year, Wu Qing, chairman of the China Securities Regulatory Commission, said in a speech at the fourth member congress of the Asset Management Association of China that it is necessary to strengthen technological empowerment, prudently and in an orderly manner explore the use of new technologies such as artificial intelligence and big data, effectively empower key scenarios such as investment and research, customer service, and internal control management, and accelerate digital and intelligent transformation.

"As an asset management company, when introducing AI tools, security is the first factor to consider," Shen Wei, chief information officer of Changxin Fund, said in a recent group interview with media including Jiemian News. For financial institutions with strong compliance requirements, he said, an organization-level AI deployment requires the institution to have systematic capabilities, whether in terms of conceptual guidance or practical issues such as data security and the boundary between humans and machines.

Shen Wei introduced that in June this year, Changxin Fund officially began using Tencent WorkBuddy. At present, it is mainly used for basic and repetitive work, while employees are freed up to engage in higher-level judgment and innovation. For example, the quantitative team uses it for factor reproduction, backtesting, and mining; the fixed income and equity teams use self-built Skills to complete decision-making chains such as financial report analysis, bond selection frameworks, and industry chain reviews.

In terms of usage results, Jiemian News learned that at present, about 60% of the daily process work of Changxin Fund's fixed income team has been taken over by Skills; the production of some operation reports and product management reports in middle-office departments has been shortened from the original 5-6 hours to 1-2 hours; and the marketing department's efficiency in producing customer companionship output on market hot spots and unusual movements has also improved significantly.

In the process of deep integration of AI and finance, financial security has also attracted industry attention and discussion. Shen Wei said that the rules for institutions using AI are completely different from personal use. Institutions must formulate refined control strategies for the scheduling and use of various models, access permissions for various types of data, and the reading permissions and scope of use of AI-generated content. Changxin Fund has different constraints for departments and personnel who come into contact with different data, and the analysis and processing of some sensitive data must call local models to prevent information leakage.

"If improving efficiency leads to data leakage or other cybersecurity incidents, the loss to the institution outweighs the gain," Shen Wei said.

Regarding the impact of artificial intelligence, Shen Wei said in response to a question from Jiemian News that as AI applications deepen, market alpha returns may become lower and lower. In the era of artificial intelligence, high-quality data has become an institution's moat. Against this background, leading institutions themselves have the advantage of historical accumulation, but if small and medium-sized institutions follow up quickly and improve their ability to use AI, they may also overtake on the curve in this wave of artificial intelligence.

Du Tingting, a quantitative researcher at Changxin Fund's Quantitative Investment Management Center, believes that artificial intelligence has lowered the threshold for obtaining data. Originally, obtaining the data and frameworks needed for quantitative work required a certain threshold, but after the emergence of AI Agents, public data sources and existing frameworks have become easier to obtain, and the speed of reproduction has been rapidly improved, which may lead to the rapid crowding and even decay of existing results.

However, Du Tingting believes that this is also an opportunity for quantitative research platforms with a foundation and accumulation. AI Agents greatly improve the efficiency of realizing ideas. Researchers and fund managers with deep understanding of the market and mature investment research frameworks can use AI Agents for rapid verification and break through the original efficiency bottlenecks.

Fan Xiaoyuan, director of Changxin Fund's Special Account Business Department, further said that the emergence of artificial intelligence has pushed all company employees to translate the implicit experience of the business into explicit rules, thereby accelerating the process of systematic development of departments and the company.

"The company actively embraces AI not just for single-point efficiency improvement, but in the hope of achieving a hierarchical leap in organizational capabilities. Through knowledge reuse and experience accumulation, we can more calmly achieve good results, increase inevitability, and reduce contingency," summarized Qin Bo, general manager of Changxin Fund.

(Source: Jiemian News) Middle of the main text Embedded advertisement