AI systems are increasingly deployed for credit assessment and investment advisory in global financial markets, yet the integrity of their inference pipelines remains insufficiently addressed by existing regulatory frameworks. This paper identifies and em



