Pakistan’s National CERT has warned that AI cyberattacks on banks and financial institutions are getting faster, larger and harder to catch, as criminals turn to AI tools, bots, botnets and agentic software. Advisory NCA-26.051126, issued on 5 November 2026, tells banks to tighten login security, watch their APIs and make sure the balance a customer sees matches the bank’s own records.
The advisory, titled “Emerging AI-Enabled Cyber Threats Targeting Financial Institutions and OT/Critical Infrastructure”, also covers power, oil and gas, water and transport operators. It follows months of government focus on banks’ ability to withstand AI-driven attacks.
- Issued by: National Cyber Emergency Response Team (National CERT), Government of Pakistan
- Advisory: NCA-26.051126, issued 5 November 2026; reported by the press on 6 and 7 November
- Who it covers: all organisations, with a focus on banks and financial institutions and on OT and critical infrastructure
- Main banking risks: automated credential attacks, disruption of online banking, attacks on APIs and payment systems, and tampering with the balances customers see
- Key measures: multi-factor authentication, defences against automated logins, privileged access management and independent transaction reconciliation
What the National CERT says about AI cyberattacks on banks
According to the advisory, attackers are using AI and automation to scout targets, find weaknesses, run credential attacks and coordinate campaigns at greater scale and speed. The Express Tribune reported that the CERT said these tools help attackers “broaden the scale, speed and sophistication of their operations.”
For banks, the advisory names core banking systems, digital banking platforms, APIs, payment systems and customer-facing apps as targets. It lists three main risks: disruption of online services, automated attacks on customer credentials, and manipulation of the account balances shown to customers. TechJuice reported the advisory’s warning that the last of these risks could cause panic “even when actual funds remain unchanged.”
The CERT also gives banks warning signs to watch for:
- a rise in automated logins, repeated authentication failures or credential stuffing
- unusual API requests, configuration changes or abnormal transactions
- mismatches between core banking records and the balances customers see
- unauthorised access to payment systems
For critical infrastructure, the advisory warns of attacks through third-party providers, connected IoT devices, industrial control systems and insecure remote access. It cites reported attacks on refrigeration and freezer systems at European supermarkets as an example of physical and commercial harm, but says AI’s involvement in those incidents has not been confirmed.
Deepfakes and AI phishing: what the advisory does not say
This advisory does not mention deepfake voice or video scams, voice cloning or AI-written phishing. Its focus is bot-driven attacks on banking systems, not tricking individual customers.
Those social engineering risks have appeared in other National CERT warnings this year. ProPakistani reported in March 2026 that a CERT advisory issued amid regional tensions listed deepfake campaigns and spear-phishing among the threats, and urged mandatory multi-factor authentication and a move away from SMS-based verification. In October, advisory NCA-18.011026 on safe use of generative AI tools told organisations to train staff on deepfakes. Advisory NCA-22.231026, issued later that month, warned of a phishing and malware campaign spreading through WhatsApp Web, WhatsApp Desktop and email.
Recommended measures for banks and fintechs
The CERT asks financial institutions to:
- review the security of core banking, APIs, third-party integrations, payment gateways and digital banking apps
- use multi-factor authentication, protection against automated logins and privileged access management
- run integrity checks between customer-facing apps and core banking records, and reconcile transactions independently
- monitor bot activity, service disruption and abnormal API use
Every organisation is asked to test its readiness for bot-driven, AI-enabled attacks, strengthen incident response and business continuity plans, keep security logs and report suspected incidents through national cybersecurity channels. The CERT lists info@pkcert.gov.pk and +92-51-9203422 as contacts.
Background: SBP Cyber Shield and earlier warnings
In February 2026 the State Bank of Pakistan launched Cyber Shield, a cyber resilience strategy under its Vision 2028 agenda. According to Dawn, all regulated entities must align their own cybersecurity programmes with the strategy, which will be rolled out in phases through 2030.
In March, the National CERT warned that hostile actors could use weak supply chains to hit power, banking and defence networks, and directed institutions to adopt zero-trust security, Business Recorder reported. On 2 May, Finance Minister Muhammad Aurangzeb met bank chiefs and their information security heads, along with the SBP governor, to review cyber preparedness. According to APP, the meeting was briefed on AI tools that can find vulnerabilities and run multi-stage attacks quickly, and the SBP and Pakistan Banks’ Association were asked to review existing frameworks and identify gaps.
The fraud numbers explain the concern. In January 2026, Minister of State for Interior Talal Chaudhry told the National Assembly that the National Cyber Crime Investigation Agency had received more than 150,000 complaints over the past year, over 81,000 of them linked to financial fraud, Profit reported. Pressure on banks is also rising as more payments go digital, a push ITInsider covered when the Prime Minister ordered a roadmap for a cashless Pakistan.
What it means for banks, fintechs and account holders
For banks, the main message is that defences built for human attackers may not keep up with bots that test thousands of passwords or API calls at once. Fintechs, wallets and payment firms that plug into banks through APIs sit inside the same risk, since the advisory names third-party integrations and payment gateways directly.
For ordinary account holders, the balance warning matters most. If an app shows an unexpected balance, the advisory’s point is that the real funds may be untouched. Checking through another official channel, such as the bank’s helpline or a branch, before reacting is sensible. Turning on every extra login protection a bank offers fits the CERT’s call for multi-factor authentication.
FAQ
When was the advisory issued? Advisory NCA-26.051126 was issued on 5 November 2026 and covered by the press on 6 and 7 November.
Has an AI attack hit a Pakistani bank? The advisory does not report any such incident. It is a warning based on threat trends.
What happens next? Banks and operators are expected to review their readiness and report suspected incidents to the National CERT. Any binding rules for banks would come from the SBP, which has not announced new measures tied to this advisory.
The National CERT’s warning on AI cyberattacks on banks shifts attention from one-off scams to automated, large-scale attacks on banking systems, and it puts the burden on banks and fintechs to prove that what customers see matches what they actually hold.
Official sources: National CERT advisory NCA-26.051126, National CERT advisories, National CERT GenAI advisory NCA-18.011026, APP
Also reported by: The Express Tribune, TechJuice, Minute Mirror, Bloom Pakistan, OECD AI Incidents Monitor
Facts and figures checked against official sources by the ITInsider desk.
Featured image: representational, AI-generated





