AI Is Raising The Pace Of Cyberattacks. Here’s How To Prepare

🔍 Read the full analysis: AI Is Raising The Pace Of Cyberattacks. Here’s How To Prepare on ThorstenMeyerAI.com

Before you orderOffer from Amazon

Get the latest gadgets delivered free with Prime

• Fast, free delivery on millions of items
• Prime Video, Amazon Music and more included
• Member-only deals all year

Start your free Prime trial →Free trial for eligible customers · Cancel anytime

As an affiliate, we earn on qualifying purchases.Hide

US agencies and cybersecurity companies have reported documented cases of attackers using AI-generated scripts and agentic software against industrial controllers and cloud services, raising concerns that automation can speed up familiar attack methods. The reports describe active activity, but also say many attempts failed; they do not establish that AI has made attacks reliably successful.

A joint advisory published on August 18 by the NSA, CISA, FBI, Department of Energy and Environmental Protection Agency says actors used AI-generated scripts disguised as monitoring tools against internet-reachable Siemens S7 programmable logic controllers, or PLCs. These industrial computers can control equipment such as pumps and production lines. The advisory says the activity involved reconnaissance and capability development, and assessed that it was likely preparation for operational effects against critical infrastructure.

In an October 21 report, Microsoft Threat Intelligence said the China-state-aligned group Storm-0940 used ARTEX, an open-source multi-agent penetration-testing framework, in activity connected to a ConnectWise Cloud Backup compromise disclosed by ConnectWise on October 16. Microsoft said the actor used credentials for the cloud backup app to attempt sign-ins to more than 2,250 customer tenants. It recorded over 151,000 failed logons in a password-spraying campaign from October 10 to 14, with a peak of 2,900 failures per minute. Microsoft said none of those attempts succeeded.

Microsoft said that in several attacks, the period from initial targeting to privileged cloud access was under three hours. It also cautioned that most targeting was unsuccessful and that an affected application or successful sign-in does not by itself show that data was accessed. ConnectWise said its existing backup archives showed no indication of compromise and that it suspended Cloud Backup, SaaS Security and Cloud Management during its investigation.

Automation Compresses Response Time

The reports point to a change in speed and scale, rather than evidence of a new, infallible kind of attacker. Scanning systems, trying credentials and exploiting known weaknesses are established activities. AI tools can help automate parts of that work, allowing activity to continue at a pace that may leave organizations less time to identify and contain an intrusion.

That matters especially where a trusted software provider or app connects to many customer environments. Microsoft’s account describes attempts to use credentials associated with a cloud backup service across thousands of tenants. Even when a campaign produces many failures, defenders must still determine which systems were reached, whether access was privileged and whether information was exposed.

Automation also adds pressure on the defensive side. OpenAI said AI helped identify critical vulnerabilities in PHP that were patched on October 21, before attackers exploited them, and forecast that AI-discovered critical vulnerabilities would “increase sharply in the coming months.” That is a forecast, not a measured rate of future discoveries. If security teams receive more actionable findings, they will need reliable asset inventories, risk-based patching and clear deadlines to respond.

Three Reports, Different Evidence

The cases described in the source material do not all carry the same evidentiary weight. The August joint advisory concerns AI-generated scripts aimed at Siemens PLCs. The agencies say actors used internet scanning services to find exposed systems and used scripts built with the open-source snap7 library. They identify manufacturing, energy, water, chemical, food and agriculture, and commercial facilities as sectors of concern. Their assessment is that the activity may be preparation for operational effects; the advisory does not say that such effects occurred.

Microsoft’s October report concerns a separate incident involving a third-party cloud backup service and Storm-0940. Microsoft describes ARTEX as capable of supporting activity from reconnaissance through data exfiltration with limited human interaction. That capability description is not proof that data was exfiltrated in this incident. Microsoft’s own caveats, along with ConnectWise’s statement about its backup archives, are important when interpreting the report.

A July 30 report from Palo Alto Networks’ Unit 42, also cited in the source material, describes a Chinese-speaking actor using the DeepSeek model with the Hermes Agent framework and Telegram. Unit 42 said the agent searched GitHub after an initial exploit attempt failed. The available material gives only part of that account, so it does not support broader conclusions about the operation’s outcome. Separately, OpenAI’s account of AI-assisted vulnerability discovery concerns defensive research, not an attack.

““This is not a theoretical risk—it is an active threat.””

— The joint advisory from the NSA, CISA, FBI, Department of Energy and EPA

Impact and Scale Remain Unsettled

The reports do not establish how often attackers are using AI across the wider threat landscape, or how much of the reported activity would have been possible without it. The evidence describes particular cases and assessments, not a comprehensive measurement of AI’s share of cyberattacks.

For the ConnectWise-related activity, Microsoft said most targeting attempts failed and that access does not necessarily mean data was taken. ConnectWise said its backup archives showed no indication of compromise. The material provided does not resolve the full scope of customer exposure or the final outcome of the company’s investigation. For the PLC activity, the agencies describe reconnaissance and likely preparation, but do not report that operational damage occurred.

A security leader quoted in the source material predicted that conditions could worsen before improving, including a possible “vulnerability tsunami” within months and a period of roughly 18 months to respond. Those are expert opinions and forecasts, not measured findings or an official timeline. The source material does not provide a named attribution or supporting methodology for those estimates.

Practical Steps for Defenders

Organizations should begin with controls that address the pathways described in the reports. The joint advisory recommends keeping an inventory of PLCs, patching firmware and preventing direct internet access to controllers. It also calls for separation between corporate IT and operational technology networks, multi-factor authentication for remote OT access, and monitoring for unusual S7 traffic. Organizations should follow the advisory for technical details and apply steps that fit their systems safely.

For cloud environments, security teams should review third-party app permissions and credentials, check sign-in logs for unusual activity and know which vendors have access to customer tenants. Microsoft’s report makes clear why organizations need a process to revoke or limit trusted app access when a supplier is under investigation. Incident response plans should be practised so teams can isolate affected accounts and systems without relying on improvised decisions.

Users can start with their email account: use a unique password, turn on phishing-resistant multi-factor authentication where available, and review connected apps and recovery settings. Organizations should set patch deadlines, keep asset inventories current and test incident response procedures. The next developments to watch are the outcome of ConnectWise’s investigation, any further advisories from government agencies, and whether future reporting provides evidence on the scale and impact of AI-assisted activity.

Key Questions

Have AI-assisted cyberattacks been confirmed?

Government agencies and Microsoft have reported specific cases involving AI-generated scripts and an agentic framework. These reports document activity, but do not show that AI makes attacks consistently successful.

Did the ConnectWise-related activity expose customer data?

The available reports do not establish that customer data was accessed. Microsoft said a successful sign-in does not necessarily mean data was accessed, and ConnectWise said existing backup archives showed no indication of compromise while it investigated.

What should companies do about exposed industrial controllers?

The joint US advisory recommends inventorying controllers, patching firmware, preventing internet access to PLCs, separating IT and OT networks, requiring multi-factor authentication for remote access and monitoring relevant traffic.

Is a surge in AI-discovered vulnerabilities certain?

No. OpenAI forecast that AI-found critical vulnerabilities would increase sharply in the coming months, but that is a forward-looking statement, not a confirmed measurement of future discoveries.

Source: ThorstenMeyerAI.com