AI’s Data Demands Uncover Governance Gaps in Financial Services

A new survey by Hitachi Vantara has highlighted significant challenges faced by the Banking, Financial Services, and Insurance sector as artificial intelligence adoption accelerates. The study, titled “The Great AI Divide: New Survey Finds Financial Services Leaders Struggle with Data Governance and Infrastructure Demands,” reveals that while 48% of financial leaders prioritise data security in AI implementations, only 36% acknowledge the critical importance of data quality, leading to compromised AI performance and return on investment .

The survey, encompassing insights from 231 global IT and business leaders, underscores a pressing concern: 84% of respondents fear that data loss due to attacks or errors would be catastrophic. Despite this, data quality remains an overlooked aspect, resulting in AI models within BFSI companies achieving accuracy merely 21% of the time, and data being available when needed only 25% of the time.

Security concerns are paramount, with 48% of participants citing data security as their top priority in AI deployment. This focus is understandable given that 36% are apprehensive about data breaches stemming from internal AI errors, and 38% express concern over the inability to recover data following ransomware attacks. Additionally, 32% are worried about AI-enabled cyberattacks leading to data breaches.

Hemant Tiwari, Managing Director and Vice President for India and SAARC Region at Hitachi Vantara, emphasised the need for a balanced approach. He stated that as AI transforms the financial sector, ensuring it operates on accurate, secure, and well-governed data is crucial. Tiwari highlighted that trust is foundational in the BFSI industry, and even minor AI-related errors can erode customer confidence and pose regulatory challenges. He advocated for building resilient data infrastructures that harmonise security with accuracy, governance, and sustainability to maintain trust and operational integrity.

The survey also reveals that despite accuracy challenges, AI adoption within BFSI is accelerating. However, many institutions are deploying AI without adequate preparation; 71% of respondents admitted to testing and iterating on live implementations, while only 4% utilise controlled sandbox environments. This approach increases vulnerabilities and underscores the need for responsible experimentation.

Alenka Grealish, co-head of Generative AI Intelligence at Celent, cautioned that while the rapid adoption of generative AI in financial services is exciting, institutions must ensure a strategic approach. She stressed the importance of balancing speed and innovation with a clear focus on security, accuracy, and ethical responsibility.