The Changing Attitude of the Financial Industry Towards Large Models: From Anxiety to Rational Exploration

Since the emergence of ChatGPT, the financial industry's attitude towards large artificial intelligence models has shifted from anxiety to rationality. Initially, there was widespread concern about being technologically lagging, and many teams were formed to work on large model-related tasks. As time has passed, financial institutions have begun to view large models more rationally, focusing on their practical application value.

Currently, several large banks have included large models in their strategic planning. According to incomplete statistics, at least 11 A-share listed banks mentioned in their latest semi-annual reports that they are exploring the application of large models. From recent trends, financial institutions are conducting deeper thinking and planning regarding large models from a strategic and top-level design perspective.

Compared to the beginning of the year, financial clients' understanding of large models has significantly improved. Some large banks, such as Agricultural Bank's ChatABC, have taken the lead in launching large model applications. Subsequently, more financial institutions have begun to focus on the practical application value of large models, rather than simply pursuing self-built models.

Due to limitations such as computing power and cost, financial institutions have adopted different strategies for applying large models. Large institutions tend to build their own enterprise-level models, while small and medium-sized institutions more frequently use public cloud APIs or private deployment services. To address computing power issues, some institutions choose to build their own computing resources, while others adopt hybrid deployment solutions.

Data governance has also become a key focus for financial institutions. More and more organizations are beginning to build data middle platforms and governance systems. Some banks have established large model data closed loops through MLOps, achieving efficient management and processing of data.

In terms of application scenarios, financial institutions generally choose to start from internal scenarios, such as smart offices and intelligent development. Code assistants and customer service assistants are relatively visible application areas. However, the application of large models in core financial business still faces challenges and requires further exploration.

Some financial institutions have begun to restructure their IT systems based on large models, adopting a layered approach that uses large models as the core to integrate traditional models. Multi-model strategies are also widely used to optimize the best results.

The application of large models has an impact on the talent structure of the financial industry. On one hand, some traditional positions face the risk of being replaced; on the other hand, there is a huge talent gap related to large models. Financial institutions are cultivating and introducing AI talents through various means to support the continuous development of large model applications.

Overall, the financial industry’s attitude towards large models has shifted from initial anxiety to rational exploration, actively seeking application paths and development models that suit its needs.