Reporting snapshot · 7 November 2026. The advisory is dated 5 November and describes risks and recommended safeguards rather than confirming any specific intrusion. It does not name an affected bank, utility or attacker, and no Pakistani institution has publicly confirmed an AI-enabled breach. The advisory also says the involvement of AI in the European incidents it cites has not been confirmed. This is a dated snapshot of official guidance, not evidence that an attack has occurred.
What happened
Pakistan’s National Cyber Emergency Response Team (NCERT) has warned banks and the operators of critical infrastructure that artificial intelligence, automated bots and agentic tools are making cyberattacks faster, broader and harder to detect. The warning is contained in an advisory dated 5 November; Dawn reported on 7 November that it had put banks and utilities on alert. The Express Tribune reported the same advisory, saying it describes a growing use of AI, botnets and automated tools in coordinated attacks on financial institutions and critical national systems.
The advisory says these capabilities may help threat actors and sophisticated criminal groups “broaden the scale, speed and sophistication of their operations”. It is not a report of a specific incident. Instead, it sets out where NCERT believes the exposure lies and what banks, utilities and other operators should do about it.
Where the risk sits
For banks, the advisory names core banking systems, digital banking platforms, payment systems and customer-facing applications as the main points of concern. It warns of disrupted online services, automated credential attacks and the manipulation of the account balances customers see in banking apps.
That last warning points to a form of harm that does not necessarily require money to move. A false balance displayed in an app can trigger panic, flood call centres and damage confidence in a bank even if the underlying ledger remains intact. The advisory treats the integrity of what customers see, and the link between that display and core banking records, as part of the security perimeter.
The infrastructure list is broader. NCERT says weaknesses at third-party service providers, connected internet-of-things devices, industrial control systems and insecure remote access could expose oil and gas facilities, water utilities, power generation and distribution networks, manufacturing plants and transport systems. “A compromise could disrupt physical processes and critical services,” the advisory warns.
It points to recent cyberattacks in Europe that affected refrigeration and freezer systems at supermarkets as an example of how a digital intrusion can have physical and commercial consequences. It also states that AI involvement in those incidents has not been confirmed, a qualification that matters for judging how much of the current threat is genuinely new.
What the advisory flags as warning signs
NCERT lists indicators that banks and infrastructure operators should watch for. On the financial side, they include frequent automated login attempts, repeated authentication failures, unusual application programming interface requests, abnormal transaction patterns and unauthorised access to payment systems.
The most specific warning is a discrepancy between core banking records and the balances shown to customers. That would suggest an attacker has found a way to alter a customer-facing layer without changing the underlying account data, or has compromised the systems that synchronise the two.
For utilities and industrial operators, the advisory points to unexpected changes in refinery, pipeline, electricity grid or power-generation parameters; unexplained alterations to pressure, temperature, cooling or heating controls; unusual remote-access sessions; and unauthorised changes to internet-of-things and operational-technology configurations.
What NCERT wants institutions to do
The recommendations are familiar security controls applied to the specific systems the advisory names. Banks are told to review the security of core banking systems, APIs, third-party integrations, payment gateways and digital banking applications. NCERT recommends multi-factor authentication, safeguards against automated login attempts and integrity checks to ensure that customer-facing applications accurately reflect core banking records.
It also asks financial institutions to monitor bot-driven login attempts, service disruptions and abnormal API activity. For utility providers, it recommends restricting direct internet exposure and remote access, reviewing the security of connected devices and industrial control systems, and strengthening the separation between operational-technology and conventional IT networks. Continuous monitoring, regular security updates and backups for essential operations are listed as baseline measures.
The advisory closes by asking all organisations to assess their preparedness for AI-driven, bot-based attacks and to maintain incident logs and incident-response, business-continuity and security-monitoring arrangements.
Why it matters
Pakistan’s banks now offer mobile apps, payment gateways and online portals for routine transactions, which is the surface the advisory describes. The same digital shift has reached electricity distribution, gas supply and transport systems, where operational technology is increasingly connected to conventional IT networks.
The advisory follows a year in which Pakistani authorities have focused on cybercrime from the enforcement side. The National Cyber Crime Investigation Agency’s raid on a Karachi call centre this month led to 273 arrests, as KhabarWire reported. The government has also moved to restrict officials from putting sensitive material into public AI tools under its cyber security handbook, a measure covered here. NCERT’s own role in registering cybersecurity consultants is set out in earlier KhabarWire reporting.
The new advisory shifts attention from data theft and fraud investigations to the resilience of systems that customers and the wider economy depend on. Its warning about displayed balances is particularly relevant to retail customers, because a convincing but false account figure could spread faster than a bank can explain the true position.
What is still uncertain
The advisory does not identify any Pakistani institution as a victim, name a threat actor or give a timeline for an attack. It also does not say whether any bank or utility has detected the indicators it lists. That leaves open the question of whether the guidance is precautionary or responds to intelligence about a specific campaign.
The AI label is also broader than it may appear. Automated credential attacks, botnets and denial-of-service campaigns long predate the recent generation of AI tools. The advisory’s concern is that AI can make those methods cheaper and faster to run, and can help attackers find weaknesses across more targets. Whether that has translated into a measurable increase in successful attacks on Pakistani systems is not established in the document.
The European supermarket example carries its own caveat: the advisory says AI involvement there has not been confirmed, so it illustrates the physical consequences of connected-system attacks rather than proving an AI-driven campaign. Readers should treat the advisory as a statement of official risk assessment and recommended practice, not as confirmation that an AI-orchestrated attack has already struck Pakistan’s financial or energy systems.
