Are Your Defenses Ready For AI-Accelerated Cyberattacks?

🔍 Read the full analysis: Are Your Defenses Ready For AI-Accelerated Cyberattacks? on ThorstenMeyerAI.com

Buying for a business?Offer from Amazon

Get business pricing on office and shipping supplies

• Business-only prices and quantity discounts
• Tax-exempt purchasing
• Multiple users, one account, clear invoices

Create a free Business account →Free to register

As an affiliate, we earn on qualifying purchases.Hide

US agencies and Microsoft have described documented cases of AI being used in cyberattack operations, including scripts aimed at industrial controllers and an agentic framework used in attacks involving a compromised backup app. The reports matter to infrastructure operators and cloud customers because they describe attackers automating familiar steps—not evidence of an all-powerful AI “super-hacker”—while leaving open whether affected organizations suffered data theft or operational damage.

A joint advisory published August 18 by the NSA, CISA, FBI, Department of Energy and Environmental Protection Agency says actors used AI-generated exploitation scripts against internet-reachable Siemens S7 programmable logic controllers, or PLCs. These industrial computers help run equipment and processes. The advisory says attackers found exposed devices through internet scanning services and disguised scripts built using the open-source snap7 library as monitoring tools. It names models from the S7-200 through S7-1500, including safety controllers, and says activity targeted sectors such as manufacturing, energy, water, chemicals, and food and agriculture.

The agencies assess the activity as likely preparation for operational effects against critical infrastructure. Their advisory calls it an active threat, but that assessment is not confirmation that the attackers caused physical disruption. It recommends security measures including taking PLCs off the public internet, patching firmware, blocking TCP port 102 at network perimeters, separating information technology and operational technology networks, and requiring multifactor authentication for remote OT access.

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

Microsoft reported that in several attacks the time from initial targeting to privileged cloud access was under three hours. It also stressed that most targeting was unsuccessful and that an affected app or successful sign-in does not necessarily mean data was accessed. ConnectWise said existing backup archives showed no indication of compromise and said it suspended Cloud Backup, SaaS Security and Cloud Management while investigating.

Automation Raises the Pace of Attacks

The reports point to a practical concern for defenders: automation can compress the time available to detect and contain an intrusion. Reconnaissance, password spraying, and attempts to adapt or run exploits are established attack methods. AI-enabled tools may let operators conduct those steps at greater speed or scale, but the cited cases do not show that AI makes every attempt effective.

For industrial operators, an exposed PLC can present risks beyond stolen files because controllers are connected to physical processes. The agencies’ recommendations focus on limiting network exposure and controlling remote access. For cloud customers, the ConnectWise case highlights how a trusted third-party app can provide a path to many customer environments. Organizations may need to review which apps hold access, what permissions they have, and how quickly unusual activity is noticed.

There is also pressure on defenders from the other direction. OpenAI said AI had identified critical PHP vulnerabilities that were patched on October 21 before attackers did, and forecast that similar findings would “increase sharply in the coming months.” That is a prediction, not a measured count of future flaws. If the pace of vulnerability discovery rises, organizations could face more work prioritizing and applying fixes; the supplied material does not quantify how many additional vulnerabilities will appear or how quickly they will be exploited.

Three Reports, Different Evidence

The reports describe distinct activities and should not be treated as a single campaign. The government advisory concerns AI-generated scripts targeting Siemens PLCs. Microsoft’s report concerns the use of ARTEX in activity linked to the ConnectWise Cloud Backup compromise and attempted access to customer tenants. The source material also cites a July 30 report by Palo Alto Networks’ Unit 42 about an actor using the DeepSeek model with the Hermes Agent framework, but the provided account cuts off before describing the agent’s actions or results. No further details about that case can be confirmed from the material supplied here.

Microsoft says Storm-0940 has been active since 2021 and describes the actor as China-state aligned. Those are Microsoft’s assessments. Its report also says ARTEX can operate from reconnaissance through data exfiltration with limited human interaction; that is a description of the framework’s capabilities, not proof that every stage occurred in this incident.

The source material includes a security leader’s forecast that conditions could worsen before improving, including a “vulnerability tsunami” within months and an estimate of roughly 18 months to prepare. This is an individual expert opinion, not a measured trend or official forecast. It should be distinguished from the dated government and company reports documenting specific activity.

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

— Joint advisory by the NSA, CISA, FBI, Department of Energy and EPA

Impact and Scale Remain Unclear

The reports do not establish that the PLC activity caused operational disruption, equipment damage, or physical harm. The supplied account also does not specify how many individual controllers were accessed or what happened after the reported targeting. Agencies’ assessment that the activity may be preparation for operational effects remains an assessment, not confirmation that such effects occurred.

For the ConnectWise incident, Microsoft’s figures include many failed login attempts, and the company said most targeting was unsuccessful. Microsoft cautioned that a successful sign-in or affected app does not necessarily mean data was accessed; ConnectWise said existing backup archives showed no indication of compromise. The supplied material does not resolve whether any customer data was viewed, copied, or otherwise affected, nor does it provide a final investigation finding.

More broadly, these cases show AI tools appearing in attack workflows, but they do not establish how common such use is across cybercrime or state-linked operations. The predicted rise in AI-found vulnerabilities and the “vulnerability tsunami” estimate are forecasts, not verified outcomes.

Priorities for Operators and Customers

Organizations can begin with controls directly tied to the reported risks: maintain an inventory of internet-facing systems and industrial controllers, patch firmware, keep PLCs off public networks, and separate OT from corporate IT. The agencies also recommend reviewing remote access, using multifactor authentication, monitoring industrial protocol traffic, and alerting on controller writes outside approved change windows.

Cloud customers should review third-party applications and the permissions they hold, including whether an app needs broad tenant access. They should monitor sign-in activity for password-spraying patterns and have a response plan for suspending a trusted integration if its credentials or service are compromised. These actions address known exposure and access paths; the reports do not suggest that buying an AI security tool alone would resolve them.

For individuals, securing the primary email account is a useful first step: use a unique password, turn on multifactor authentication, and review account recovery methods and connected apps. The next concrete developments are the outcome of ConnectWise’s investigation and any follow-up advisories or technical details from the agencies and companies. Until those appear, the confirmed record supports vigilance about automation and access—not claims that every organization has been breached.

Key Questions

Have AI-assisted cyberattacks been documented?

Yes. A joint US agency advisory says actors used AI-generated scripts against Siemens S7 controllers, and Microsoft says Storm-0940 used the ARTEX agentic framework in activity linked to attacks involving ConnectWise Cloud Backup.

Does the ConnectWise report confirm that customer data was stolen?

No. Microsoft says a successful sign-in or affected app does not necessarily mean data was accessed. ConnectWise said existing backup archives showed no indication of compromise, and its investigation was ongoing in the supplied material.

Did the password-spraying attempts succeed?

Microsoft reported more than 151,000 failed logons from October 10 to 14, with a peak of 2,900 failures per minute, and said none of those attempts succeeded. It separately described successful access in several attacks, so the failed password-spraying campaign should not be conflated with all activity in the report.

What should industrial operators do now?

The joint advisory recommends keeping PLCs off the public internet, patching firmware, blocking TCP port 102 at the perimeter, separating IT and OT networks, requiring multifactor authentication for remote OT access, and monitoring for unapproved controller writes.

Is a surge in AI-found vulnerabilities confirmed?

No. OpenAI forecast that similar AI-found vulnerabilities would increase sharply in coming months, but that is a forecast. The source material provides no measured future count or baseline for such an increase.

Source: ThorstenMeyerAI.com