The rapid advance of generative AI is reshaping the threat landscape for banks and financial institutions, forcing banks to confront a new convergence of fraud risk, regulatory scrutiny and digital resilience expectations.
Fresh warnings from the World Economic Forum (WEF), taking place from January 19 to January 23 in Davos, underline that AI-driven cybercrime is no longer a peripheral concern.
Instead, it should be seen as a core software safety issue with direct implications for financial stability, digital resilience and regulatory compliance, experts at the annual summit in Switzerland have warned.
According to the WEF, AI-enabled fraud has now overtaken ransomware as the primary cyber risk facing business leaders.
At the same time, regulators are increasingly focused on whether AI systems used in financial services can be reliably tested, explained and controlled across their lifecycle, a shift that places QA and software testing at the centre of AI governance.
A WEF report that was shared with QA Financial shows that 73% of surveyed CEOs, or someone in their professional or corporate network, had been affected by cyber-enabled fraud in 2025.
Phishing, vishing and smishing attacks were reported by 62% of respondents, while 37% encountered invoice or payment fraud and 32% reported identity theft.
The survey highlights a decisive change in executive priorities, with AI vulnerabilities and cyber-enabled fraud now eclipsing ransomware as the dominant concern, a signal that the attack surface has moved beyond infrastructure into software behaviour, identity systems and decision-making logic.
From a QA perspective, that shift matters, delegates in Davos argued during a panel discussion. AI-driven fraud exploits weaknesses not only in perimeter security, but in how systems authenticate users, validate transactions, detect anomalies and respond under stress. Those capabilities depend heavily on how software is tested and monitored in real-world conditions.
AI lowers the barrier for attacks
Konstantin Levinzon, co-founder of Planet VPN, warned that generative AI is accelerating both the scale and credibility of cyber scams, with implications for businesses and consumers alike.
“As businesses face challenges in protecting their networks, individual consumers are also seeing an increase in personal cybersecurity risks,” Levinzon explained.
“Recent developments in generative AI are lowering the barriers to executing all kinds of attacks, while at the same time increasing their sophistication and making them appear more credible,” he said.
Recent data from the US Federal Trade Commission shows that consumers reported $12.5 billion in fraud losses in 2024, a 25% year-on-year increase.
Levinzon predicts that this figure could rise further as AI becomes more widely adopted by criminal networks.
For banks, that trend raises questions about how fraud-detection models are tested, how quickly they adapt to new attack patterns, and whether QA processes can surface failure modes before losses occur. Static test cases and historical data sets are increasingly inadequate in an environment where attack techniques evolve continuously.
Levinzon also points to the role of AI in expanding the geographic and linguistic reach of scams.
“Criminal networks that previously focused on a limited range of languages can now target populations all over the world with local languages,” he said.
“This expansion also speeds up the spread of AI-driven disinformation and makes it harder for platforms and regulators to protect users from coordinated manipulation,” the Greece-based entrepreneur added.
The WEF noted that generative AI amplifies digital safety risks for vulnerable groups, including children and women, who are increasingly targeted by impersonation and synthetic image abuse.
For financial institutions, that adds pressure to ensure customer-facing systems behave safely under edge cases, social engineering attempts and identity spoofing scenarios, all of which are testing challenges as much as security ones.
Skills shortages meet automation risk
The WEF report also highlighted this week a persistent shortage of cybersecurity expertise, with 33% of firms in Europe and Central Asia and 35% in North America reporting skills gaps. In parts of Latin America and Africa, up to 70% of companies face shortages.
Levinzon argued that AI tools can help address those gaps, but only if implemented with care.
“However, if implemented poorly, AI can introduce new risks of misconfiguration, biased decision-making, over-reliance on automation, and susceptibility to adversarial manipulation,” he stated.
For QA teams, this reinforces the need for rigorous validation, configuration testing and continuous monitoring of AI-driven systems. Automation may ease operational pressure, but without strong testing controls it can amplify risk rather than reduce it.
AI governance in finance
Alongside its warnings on fraud, the WEF delivered a parallel message to financial regulators and institutions: the global financial sector should strengthen AI risk controls rapidly.
A separate WEF report, Artificial Intelligence in Financial Services, released as leaders gather in Davos, makes clear that regulatory expectations around AI are no longer abstract.
Supervisors are increasingly focused on how firms can demonstrate that AI systems are reliable, explainable, resilient and compliant throughout their lifecycle.
In practice, those gaps manifest as testing failures: unclear validation strategies, limited visibility into model behaviour, and inadequate monitoring once systems are live.
For QA and software testing teams, this is likely to represent a fundamental shift. Testing is no longer something that happens after development, but a core mechanism through which AI risk is controlled and evidenced.
The report highlighted persistent gaps between AI ambition and operational readiness, particularly where firms struggle to demonstrate oversight and accountability for complex models.
Traditional functional testing is no longer sufficient when AI influences credit decisions, fraud detection, pricing, customer interactions or market activity.
Regulators are implicitly asking how institutions test not just whether systems work, but whether they behave appropriately under stress, edge cases and changing data conditions.
Compliance a major challenge
The WEF also pointed to uneven regulatory clarity across jurisdictions, with AI rules emerging alongside existing conduct, risk and resilience frameworks.
That uncertainty does not reduce accountability. Instead, it increases pressure on firms to demonstrate control through evidence.
For software testing teams, compliance becomes a testability challenge. AI systems must be designed so their outputs, decision logic and failure modes can be tested, audited and explained.
Where models cannot be meaningfully tested or monitored, they become regulatory liabilities regardless of their performance benefits.
The report stressed that governance frameworks alone are insufficient without technical mechanisms to support them.
Testing pipelines, validation environments, controlled data sets and continuous monitoring are what turn responsible-AI principles into something regulators can accept in practice.
This shifts QA from a delivery function to a strategic one. Testing becomes ongoing, adaptive and tightly integrated with production systems, particularly as AI models evolve with new data.
The WEF’s emphasis on resilience and systemic risk implicitly calls for stress testing AI systems under conditions that mirror real-world volatility, adversarial behaviour and unexpected inputs.
In that context, QA teams intersect directly with enterprise risk management. AI failures are no longer isolated defects; they can trigger conduct breaches, financial losses and reputational damage.
Taken together, the WEF’s warnings on AI-driven fraud and AI governance reinforce a single message for banks and financial services firms: digital resilience now depends as much on how software is tested as on how it is built.
As regulators scrutinise AI deployment and cyber risks intensify, QA teams are emerging as one of the sector’s most important lines of defence, responsible not just for quality, but for trust, safety and systemic stability in an AI-driven financial system.
COMING IN 2026
Why not become a QA Financial subscriber?
It’s entirely FREE
* Receive our weekly newsletter every Wednesday * Get priority invitations to our Forum events *
REGULATION & COMPLIANCE
Looking for more news on regulations and compliance requirements driving developments in software quality engineering at financial firms? Visit our dedicated Regulation & Compliance page here.
READ MORE
- Inside banking’s shift to smarter QA to tackle complexity and risk
- SmartBear CPTO on AI in banking QA: ‘Impressive metrics but no critical scenarios’
- Banks push beyond traditional QA as resilience testing gains ground
- Banking QA professionals warn AI still doesn’t know ‘where the bodies are buried’
- RECAP: The QA Financial Healthcare & Insurance Forum Philadelphia 2026
QA FINANCIAL PODCASTS
LISTEN TO OUR EXCLUSIVE CONVERSATIONS
WATCH NOW
