This week’s dominant theme is the convergence of AI-driven threat and defence across financial services: agentic AI is simultaneously industrialising fraud at unprecedented scale while banks and insurers race to deploy the same technology as a countermeasure. Alongside this, the standalone AI liability insurance market is crystallising at Lloyd’s, the EBA is clarifying how the EU AI Act sits alongside existing banking law, and a major DIFC report warns traditional banks face $170bn in profit erosion if they fail to respond to AI-native challengers.
Top story: BCG warns agentic AI could slash the cost of running financial scams by 90% within two years, triggering a fraud surge that banks’ current defences are not built to handle.
Agentic AI Could Cut Fraud Costs 90%, BCG Warns Banks
BCG · Risk
A BCG report published on 11 June warns that agentic AI could reduce the cost of executing financial scams by 90% within two years, leading to a twofold or greater surge in attack volume. GenAI is already enabling synthetic identities, cloned voices, and fake documents at scale — but the real inflection point comes when agentic systems can run fraud end-to-end without human intervention. BCG urges banks to redesign prevention, detection, and response operating models now, before adversaries outpace them.
https://www.bcg.com/publications/2026/how-agentic-ai-will-industrialize-financial-scams
Traditional Banks Face $170bn Profit Hit From AI Challengers
Gulf News · Strategy
The DIFC’s 2026 Future of Finance report warns that global banks could face $170 billion in profit erosion by 2030 as AI-native challenger banks set new standards for speed, personalisation, and cost efficiency. The report argues that competitive advantage will hinge on how well banks clean, connect, and deploy proprietary real-time data — not on size or legacy. AI-first cloud infrastructure is rapidly becoming table stakes, not a differentiator.
EBA: EU AI Act Complements Banking Law, No New Guidelines Needed
Bird & Bird · Regulation
The European Banking Authority has confirmed it finds no significant contradictions between the EU AI Act and existing EU banking and payments regulations, stating the two frameworks are complementary rather than duplicative. The EBA will not issue new standalone guidelines but will spend 2026–2027 promoting a common supervisory approach across member states ahead of the August 2026 high-risk AI compliance deadline. For financial institutions, this means integrating AI Act obligations into existing DORA and risk management frameworks rather than building entirely separate compliance programmes.
Lloyd’s-Backed Testudo Shapes Standalone AI Liability Market
The Insurer / Program Manager · Finance
Testudo, a Lloyd’s Lab-backed MGA supported by LocalGlobe and Goldman Sachs partners, began underwriting AI liability policies for US mid-market enterprises in early 2026, targeting vendors and deployers of AI systems as traditional insurers quietly add AI exclusions to E&O, GL, and D&O lines. The platform analyses real-time AI litigation, regulatory incidents, and risk signals to quantify exposures that conventional underwriting misses. Executives say the market is in its infancy but that rising AI deployment failures and the shift to agentic systems will drive strong future demand.
FCA Expands ‘Supercharged Sandbox’ for AI-Driven Fintech Products
Global Policy Watch · Regulation
The UK’s Financial Conduct Authority confirmed in its 2026/27 work programme the expansion of its Supercharged Sandbox to a new cohort of firms, giving participants access to high-quality synthetic data to test AI-driven financial products in a controlled environment. The move signals the FCA’s preference for enabling live experimentation over prescriptive AI-specific rules, even as parliamentary pressure mounts following the Treasury Committee’s January 2026 warning that a ‘wait-and-see’ approach risks consumer harm. The FCA has also flagged emerging perimeter risks from general-purpose AI tools offering financial advice outside the regulatory boundary.