arXiv:2605.23955v1 Announce Type: cross Abstract: Deploying machine learning in regulated financial environments -- credit risk, fraud detection, and anti-money laundering -- exposes critical vulnerabilities in algorithmic reproducibility. While early financial ML addressed statistical challenges such as backtest overfitting, deep neural networks and Generative AI have introduced mechanical nondeterminism rooted in hardware and architecture. This survey provides a systems perspective on reproducibility failures across three modalities now dominant in financial AI: tabular models (post-hoc expl
The proliferation of complex AI models like deep neural networks and Generative AI in financial services, coupled with increasing regulatory scrutiny, is forcing a re-evaluation of algorithmic reproducibility and auditability.
Non-determinism in financial AI systems poses significant risks to regulatory compliance, trust, and the stability of critical financial functions like fraud detection and credit risk, potentially leading to systemic issues.
The focus in financial AI development will shift from purely accuracy-driven metrics to encompassing auditability and reproducibility, directly influencing model design and deployment strategies in regulated sectors.
- · Financial Regulators
- · Firms specializing in AI explainability and audit tools
- · Robust AI system architects
- · Traditional financial institutions leveraging transparent AI
- · Black-box AI model developers
- · Financial institutions with opaque legacy AI systems
- · Unregulated AI solutions in finance
- · High-frequency trading firms reliant on unexplainable models
Increased investment in explainable AI (XAI) and reproducible machine learning practices within the finance industry.
New regulatory frameworks and compliance standards specifically addressing the determinism and auditability of AI used in critical financial applications.
A potential bifurcation in AI adoption, where regulated financial applications prioritize transparency and auditability over bleeding-edge performance, while other sectors continue to push raw performance boundaries.
This signal links to a primary source. Continuum Brief monitors and indexes it as part of the live intelligence stream — we do not republish source content.
Read at arXiv cs.LG