AI in Banking and Insurance Operations — 2026-09-26
This week, supervisory scrutiny of agentic AI intensified as US examiners gained a template for policing AI and Wall Street's agents moved from advising to acting. In Korea, insurers face criticism for thin AI governance (only 7.5% have formal AI control bodies) while a joint government-industry forum targets AI- and deepfake-enabled insurance fraud. Global banks, including NatWest and Bank of America, are warning that AI shopping agents open new fraud exposure.
AI in Banking and Insurance Operations — 2026-09-26
Top developments
Wall Street's AI agents "have stopped advising and started working"
A report this week finds JPMorgan and Morgan Stanley are moving beyond chatbots to autonomous software that acts on its own, and notes central banks are already asking who can stop it. The shift matters for call-centre bots and AI advisers: deployment is migrating from recommendation engines to systems that execute transactions, raising new control questions for risk teams.
American Banker: examiners already have a template for regulating AI
In an opinion piece published this week, American Banker argues that recent federal guidance stopped short of AI-specific rules, but bank examiners already have tools to ensure AI systems "operate safely" — pointing to existing model-risk examination practice. The debate follows the CSBS framework and revised interagency model risk guidance, signalling that AI governance conversations are moving from policy papers to exam rooms.
AI deployment moves into core workflows — with a value test
Finance Monthly reports this week that financial services face "a new AI value test": banks and fintechs are pivoting from pilots to core workflows and measurable business outcomes rather than experimentation. This matters for fraud models and underwriting engines, where boards are increasingly demanding quantified ROI before scaling.
Global banks warn on AI shopping-agent fraud
Global Economic reported (September 23) that six major global banks including NatWest and Bank of America are urging transaction disclosure and consumer protection measures as AI shopping agents that select and pay for goods raise fraud and data-leak risks. Payment-authorised agent flows could become a new attack surface that fraud models must cover.
Local view
Korean media this week focused heavily on insurer AI governance gaps. Herald Economy reported that only 7.5% of insurers have established formal AI governance ("통제기구") structures, even as authorities in June strengthened financial-sector AI discipline into firm-wide risk management — and recommended tighter controls on consumer-facing tasks like claims adjudication. DigitalToday covered the US state-examiner AI guidance for Korean readers, noting it covers generative and agentic AI excluded from federal model-risk guidance. Straight News (September 23) reported insurers are split between excluding generative-AI risks from existing policies and offering standalone coverage, as wrong AI answers, copyright and privacy harms cascade across multiple corporates — a live underwriting decision for Korean carriers. Chosun Ilbo reported (September 23) the Financial Supervisory Service is inspecting insurers' actuarial-organisation controls against profit inflation.
Context & numbers
- Lloyds Banking Group expects its 2026 agentic AI deployment to add £100 million in value through automated fraud investigations; Swiss Re's ClaimsGenAI generated over 1,000 fraud alerts in its first year.
- JPMorgan runs LLM Suite for 200,000+ employees within an $18B 2026 tech budget; BlackRock's Aladdin applies AI risk analytics to ~$21T in assets.
- AI-enabled fraud losses in banking are projected to reach $40 billion by 2027.
On the radar
- Korea's revised Financial AI Guideline and FS's AI Risk Management Framework are rolling out to insurers — EBN has begun a series on the seven core principles; watch supervisory follow-up in claims processing.
- Seoul Finance is tracking how bank staff co-work with AI and the associated risk-management buildout post-deployment.
- Korea Post is running a series on jobs disappearing via AI in banking and insurance underwriting — a leading indicator of domestic headcount pressure.
- Wells Fargo risk-management succession news surfaced in this week's fintech briefings — worth watching for how its AI risk leadership lands.
This content was collected, curated, and summarized entirely by AI — including how and what to gather. It may contain inaccuracies. Crew does not guarantee the accuracy of any information presented here. Always verify facts on your own before acting on them. Crew assumes no legal liability for any consequences arising from reliance on this content.
Powered by
