Saiprakash Kodela Develops Trustworthy Banking AI as UK Fraud Losses Hit £1.28 Billion
Updated
Updated · aijourn.com · Jul 10
Saiprakash Kodela Develops Trustworthy Banking AI as UK Fraud Losses Hit £1.28 Billion
1 articles · Updated · aijourn.com · Jul 10
Summary
Saiprakash Kodela is advancing AI systems for banking that emphasize explainability, auditability, privacy and zero-trust controls rather than fraud detection alone.
£1.28 billion in UK fraud losses in 2025, more than $20 billion in reported U.S. cyber-enabled crime losses, and Deloitte’s $11.5 billion 2027 email-fraud forecast underscore why banks need stronger defenses.
His work spans U.S. patent applications, a German utility model and UK registered designs covering explainable multi-agent fraud detection, open-banking monitoring, privacy-preserving cross-bank analytics and zero-trust AI orchestration.
That research builds on enterprise banking infrastructure work including trading-system database migration, secure credential retrieval, monitoring repair automation and audit-trail preservation in production environments.
The broader pitch is that as real-time payments, open banking and autonomous AI expand finance’s attack surface, banks will need control architectures that make AI trustworthy at scale.
As criminals weaponize AI, can our financial defenses ever truly catch up?
When an AI wrongly freezes your bank account, who is ultimately held responsible?
AI Arms Race Fuels UK Fraud Crisis: £354 Million Reimbursed to Victims in 2025, Demanding New Defenses
Overview
The United Kingdom is facing a major and evolving fraud crisis, with substantial financial losses and a threat landscape that is rapidly changing due to artificial intelligence. In 2025, banks reimbursed £354.3 million to victims of Authorised Push Payment fraud, which was 61% of total losses, and firms prevented £1.68 billion in unauthorised fraud. This progress is partly due to the UK APP fraud reimbursement regime. However, the rise of AI has made fraud more sophisticated, challenging traditional detection methods and requiring new, innovative approaches to keep up with increasingly advanced threats.