The U.S. Senate Committee on Banking, Housing, and Urban Affairs held a subcommittee hearing on July 30, 2025, focusing on the role of artificial intelligence (AI) in capital and insurance markets. The meeting emphasized the importance of establishing guardrails to manage risks associated with AI while promoting transparency and trust within the financial sector.

Key discussions highlighted the need for transparency in the data used to train large language models (LLMs). Experts expressed concerns that many developers do not fully disclose the data sources, which can lead to significant issues, especially in regulated industries like finance. Financial leaders have voiced a strong desire to understand the data underpinning the AI models they utilize, as this knowledge is crucial for building trust and ensuring compliance with regulatory standards.

The hearing underscored the urgency of adopting AI technologies to gain competitive advantages for both the economy and consumers. However, it was noted that responsible AI governance is essential to expedite this adoption. The principles of open, trusted, and secure AI were highlighted as foundational to IBM's approach, advocating for open-source AI to foster collaboration and innovation while ensuring that smaller firms can compete effectively.

Three key recommendations were proposed for policymakers:

1. Support open AI ecosystems to encourage collaboration between academia, industry, and government.
2. Implement use case-based risk frameworks to regulate AI applications according to their specific contexts.
3. Establish transparency requirements that allow enterprises to disclose the data used for training models and the decisions influenced by these models.

The discussions concluded with a consensus that while generative AI presents risks, its potential social and economic benefits far outweigh these concerns. The focus moving forward will be on how to deploy AI thoughtfully and responsibly, ensuring that innovation is balanced with necessary safeguards.