How AI Changes The Time Companies Have To Stop Cyberattacks

🔍 Read the full analysis: How AI Changes The Time Companies Have To Stop Cyberattacks 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 researchers have reported attackers using AI-generated scripts against industrial controllers and an AI agent in a cloud supply-chain campaign, adding evidence that automation can shorten the time defenders have to respond. Microsoft said that in several incidents linked to Storm-0940, attackers moved from initial targeting to privileged cloud access in under three hours, while cautioning that most targeting attempts failed and access did not necessarily mean data was accessed.

A joint advisory published August 18 by the NSA, CISA, FBI, Department of Energy and Environmental Protection Agency said actors were conducting reconnaissance and developing capabilities against Siemens S7 programmable logic controllers. The agencies said attackers found internet-exposed controllers through scanning services and used AI-generated exploitation scripts disguised as monitoring tools, built with the open-source snap7 library. The activity targeted industrial sectors including manufacturing, energy, water, chemicals, food and agriculture, and commercial facilities.

Microsoft Threat Intelligence reported October 21 that Storm-0940, which it describes as aligned with the Chinese state and active since 2021, used the open-source ARTEX multi-agent penetration-testing framework after the ConnectWise Cloud Backup compromise. Microsoft said ARTEX was used to attempt sign-ins to more than 2,250 tenants after credentials for the backup application were obtained. Between October 10 and 14, Microsoft logged more than 151,000 failed sign-ins in a password-spraying campaign, with a peak of 2,900 failures per minute; Microsoft said none of those attempts succeeded.

Microsoft’s report also says that in several attacks the interval from initial targeting to privileged cloud access was under three hours. It stressed that most targeting was unsuccessful and that an affected application or successful sign-in does not by itself show that information 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.

Shorter Windows for Cloud Defense

The reports point to a practical change in the pace of familiar attack steps, rather than proof of a new kind of autonomous attacker. Scanning, password spraying, exploit adaptation and abuse of trusted applications are established techniques. Automating parts of that work can let attackers test more targets and keep operating without pauses, leaving security teams less time to identify suspicious activity and limit access.

That matters particularly when attackers exploit trusted software or exposed operational technology. A compromised application can carry access into customer environments, while an internet-reachable industrial controller may affect systems that run physical processes. The agency advisory recommends measures such as asset inventories, firmware updates, keeping controllers off the public internet, separating IT and operational networks, and monitoring industrial traffic. These are established defenses, but the reports underline the cost of leaving them incomplete.

AI can also increase the work defenders face. OpenAI said AI had identified critical vulnerabilities in PHP that were patched October 21 before attackers exploited them, and forecast that similar discoveries would rise sharply in coming months. That is a company forecast, not a measured prediction of how many flaws will be found or how quickly organizations can patch them. More vulnerability reports can help security, but they can also add pressure to prioritize and deploy fixes.

Three Reports Show the Pattern

The industrial-control advisory describes activity against Siemens S7 PLCs, computers used to control equipment and processes. The agencies assess that the activity may be preparation to cause operational effects, but that assessment is not confirmation that such effects occurred. They recommend blocking unnecessary internet exposure, applying firmware updates, requiring multifactor authentication for remote access, and separating operational technology from corporate IT.

Microsoft’s October report concerns a different path: the ConnectWise Cloud Backup supply chain. ConnectWise disclosed the compromise October 16; Microsoft later described Storm-0940’s use of ARTEX and attempts to use the compromised app’s credentials across customer tenants. Separately, Palo Alto Networks’ Unit 42 reported July 30 on an actor using a DeepSeek model with the Hermes Agent framework, directing it through Telegram. Unit 42 said the agent adapted after an exploit attempt failed. Together, the reports describe automation being added to ordinary attack workflows, not evidence that AI independently conceived every step.

In its October 21 account, Microsoft said the ARTEX framework could operate from reconnaissance through data exfiltration with limited human interaction. That is Microsoft’s description of the framework’s capabilities; it does not establish that every stage succeeded in the reported campaign. The distinction matters: demonstrated use of an agent is confirmed by the report, while the extent of any resulting data access remains a separate question.

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

— The joint US agency advisory

What the Reports Do Not Establish

The reporting does not establish how many industrial controllers were successfully exploited, whether the activity described by the agencies caused operational disruption, or how many organizations suffered confirmed data theft. The agencies describe preparation and capability development and assess that actors are likely preparing for operational effects; they do not report that those effects have occurred.

For the ConnectWise-related activity, Microsoft said a successful sign-in or affected application does not necessarily mean data was accessed. ConnectWise said its existing backup archives showed no indication of compromise, but the supplied material does not resolve the full scope or final outcome of the investigation. The number of attempted tenant sign-ins should not be read as a count of successful compromises.

A security leader quoted in the source material forecast a coming “vulnerability tsunami” and said organizations may have roughly 18 months to respond. That is an expert opinion, not a measured timetable or an established forecast. The reports also do not quantify how much faster AI made these campaigns compared with the same activity conducted without AI.

Investigations and Basic Controls

ConnectWise’s investigation and any subsequent disclosures may clarify whether customer data was accessed and the extent of the compromise. Security teams will also be watching for further agency updates on the industrial-control activity and whether reported reconnaissance develops into confirmed operational incidents.

For organizations, the recommendations in the agency advisory and incident reports point to immediate defensive priorities: inventory exposed assets, patch on defined deadlines, require phishing-resistant sign-in where available, review third-party application access, and separate operational technology from IT. Companies should also rehearse incident response so staff can rapidly disable suspicious access and assess affected systems. For individual users, securing the email account tied to password resets and other services is a practical starting point.

Whether AI changes attack outcomes at scale remains an open question. The confirmed cases show that attackers are using the tools in real operations, but the available reporting also records failed attempts and unresolved questions about impact. Further incident findings will show whether the shorter response windows described in some cases become a broader pattern.

Key Questions

What is the main development?

Government agencies and security researchers have documented attackers using AI-generated scripts and agent frameworks for parts of cyber operations, including activity against industrial controllers and cloud tenants.

Did the Storm-0940 campaign compromise more than 2,250 organizations?

No. Microsoft said the actor attempted sign-ins to more than 2,250 tenants. It also said most targeting was unsuccessful, and that a successful sign-in does not necessarily mean data was accessed.

Were industrial systems reported as disrupted?

The advisory describes reconnaissance and capability development against internet-reachable Siemens controllers and says the activity may be preparation for operational effects. The supplied reporting does not confirm resulting disruption.

What should companies prioritize?

Start with asset inventories, timely patching, strong authentication, third-party access reviews, network separation and practiced incident response. For industrial systems, the advisory also recommends preventing direct internet access and monitoring controller traffic.

Is a coming vulnerability surge confirmed?

No. OpenAI forecast that AI-found critical vulnerabilities would rise sharply in coming months. That remains a forecast; the source material does not provide a measured estimate of the increase.

Source: ThorstenMeyerAI.com