Financial regulators are increasingly scrutinizing the rapid adoption of artificial intelligence agents in banking and enterprise systems, highlighting gaps in oversight and operational control. In Australia, the Australian Prudential Regulation Authority warned that while banks and superannuation funds are actively deploying AI in areas such as fraud detection, loan processing, and customer service, many institutions still lack mature risk management frameworks. The regulator found uneven governance practices, weak monitoring of model behavior, and overreliance on external vendors without sufficient internal scrutiny. It also raised concerns about emerging cybersecurity threats, including prompt injection attacks and insecure system integrations involving non-human AI agents. Additionally, identity and access management systems were often not adapted to account for autonomous AI workflows. Industry groups such as the FIDO Alliance are now developing new authentication standards for AI-driven actions, reflecting the growing need to distinguish between human and machine-initiated decisions. Overall, regulators emphasize that stronger governance, transparency, and human oversight are essential as AI systems take on more critical operational roles.
Read more here
https://www.artificialintelligence-news.com/news/ai-agent-governance-control-gaps/
