This week’s stories cluster around two major themes: the EU Parliament’s landmark vote to delay high-risk AI deadlines for financial institutions, buying banks and insurers more than 16 months of compliance breathing room; and a sharp escalation in AI-powered fraud, with UK research showing 84% of fraud professionals say AI has materially increased the complexity of attacks. Across both themes, the pace of AI adoption in finance is clearly outrunning institutions’ readiness to govern or defend against it.
Top story: EU Parliament formally approves Digital Omnibus on AI, pushing the high-risk compliance deadline for bank and insurer AI systems to December 2027.
EU Parliament Votes to Delay Bank and Insurer AI Deadline to 2027
Dastra (EU regulatory intelligence) · Regulation
On 16 June 2026, the European Parliament voted 423 to 57 to approve the Digital Omnibus on AI, formally pushing the compliance deadline for high-risk AI systems — including credit scoring, AML tools, and insurance underwriting models — from August 2026 to December 2027. The Council is expected to formally adopt the text on 29 June, giving financial institutions across Europe a crucial 16-month reprieve. However, transparency and watermarking obligations under Article 50 remain live from August 2026, meaning compliance teams cannot fully stand down — the clock has been reset for some systems and left running for others.
https://www.dastra.eu/en/blog/digital-omnibus-on-ai-parliament-votes-deadlines-redrawn/60108
Agentic AI Is Making UK Bank Fraud ‘Exponentially’ Harder to Stop
FinTech Global · Risk
BioCatch published new research drawing on 80 UK fraud and financial crime professionals, finding that 84% believe AI has heightened the complexity of fraud and scam activity, while 73% report a rise in fraud attempts at their institutions. The report warns that agentic AI is enabling fraudsters to operate at unprecedented speed and scale, with the proportion of banks reporting year-on-year fraud loss growth rising from 59% to 76% in a single year. The firm argues that static identity checks are no longer sufficient and calls for real-time behavioural intelligence shared across the sector.
https://fintech.global/2026/06/11/agentic-ai-puts-uk-fraud-defences-under-pressure/
AI Is Hollowing Out the Mass-Affluent Wealth Management Market
Bloomberg · Strategy
McKinsey’s senior partner Debasish Patnaik told Bloomberg this week that mass-affluent clients now receive something close to private-banking quality service from AI, fundamentally undermining the value proposition of mid-tier advisers. The shift is forcing wealth management firms to reconsider who they hire and how they serve this historically profitable segment, as AI commoditises what was once bespoke advice. For wealth managers and private banks, the story signals a structural squeeze on adviser headcount and fee models in the segment sitting just below ultra-high-net-worth.
PwC: AI Reshaping PE Diligence as Deal Volume Falls 34%
PwC · Finance
PwC’s freshly published 2026 mid-year private equity outlook shows deal volume in H1 2026 fell 34% even as average deal size rose nearly fourfold, as capital concentrated in fewer, higher-conviction bets. The report flags that AI is now embedded in the investment thesis itself — sponsors must assess how exposed a target is to AI disruption, whether it has the data foundations to adopt AI, and how management plans to use AI to improve margins. For PE practitioners, AI capability has moved from a value-creation add-on to a core underwriting criterion and a determinant of exit readiness.
https://www.pwc.com/us/en/industries/financial-services/library/private-equity-deals-outlook.html
EU Parliament Confirms National Regulators Will Supervise AI at Financial Institutions
Council of the EU · Regulation
A provision tucked inside the Digital Omnibus on AI agreement clarifies that national competent authorities — not the EU AI Office — will retain supervisory responsibility over AI systems at financial institutions, even where those systems are built on general-purpose AI models. This is a significant structural decision: it means that UK-equivalent national supervisors across EU member states will lead enforcement for bank and insurer AI, rather than a centralised EU body. For compliance teams at European banks, this signals that regulatory scrutiny will come via their existing domestic financial regulator, aligning AI Act oversight with DORA and MiFID II supervisory channels.