🔍 Read the full analysis: Can Companies And Users Keep Up With AI-Fueled Cyberattacks? on ThorstenMeyerAI.com
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
TL;DR
US agencies say attackers have used AI-generated scripts against internet-reachable industrial controllers, and Microsoft reports that an AI agent was used in attacks involving a compromised cloud backup app. Microsoft also says most targeting attempts failed and that successful access does not by itself prove data theft. The cases show AI can speed routine attack work, while the scale of future threats remains uncertain.
US agencies and Microsoft have reported AI-assisted cyber activity against industrial controllers and cloud services, including an attack in which an agentic framework was used to target Microsoft 365 tenants through a compromised backup app. The reports show that software can speed up familiar attack steps, but they do not establish that AI makes every attack successful: Microsoft said most targeting failed, and that a successful sign-in does not necessarily mean data was accessed.
A joint advisory issued Aug. 18 by the NSA, CISA, FBI, Department of Energy and EPA says threat actors used AI-generated scripts against internet-reachable Siemens S7 programmable logic controllers, or PLCs. These industrial computers help operate equipment in sectors including energy, manufacturing and water. The agencies said the activity included scanning for exposed controllers and scripts disguised as monitoring tools, and assessed it as likely preparation for possible operational effects. The advisory called the threat active, not theoretical.
On Oct. 21, Microsoft Threat Intelligence described activity involving Storm-0940, which Microsoft identifies as a China-state-aligned actor active since 2021. Microsoft said the actor used ARTEX, an open-source multi-agent penetration-testing framework, after obtaining credentials for the ConnectWise Cloud Backup app. The company reported attempts to sign in to more than 2,250 tenants and more than 151,000 failed logons from Oct. 10 to 14, with a peak of 2,900 failures per minute. None of those password-spray attempts succeeded, Microsoft said.
Microsoft said that in several attacks, the time 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 necessarily show that data was accessed. ConnectWise disclosed the compromise on Oct. 16 and said existing backup archives showed no indication of compromise; it suspended Cloud Backup, SaaS Security and Cloud Management while investigating. Separately, OpenAI said AI had found critical PHP vulnerabilities patched on Oct. 21 and forecast that AI-found critical vulnerabilities would “increase sharply in the coming months.”
Infographic 01 · The record so far
AI is in the attacker’s toolchain. On the record.
Six documented events in twelve weeks. Each label says what the source itself states about AI. Figures are the vendors’ and agencies’ own.
- 30 Jul 2026Unit 42: AI on attackDeepSeek inside the Hermes Agent framework picks and switches targets on its own. Limited impact: across autonomous and manual attempts, data theft from 3 NetScaler systems confirmed. Single-vendor report.
- 18 Aug 2026NSA, CISA, FBI, DOE, EPA: AI on attackAI-generated exploitation scripts, disguised as monitoring tools, against Siemens S7 PLCs in US critical infrastructure. “Not a theoretical risk.”
- 10–14 OctStorm-0940 password spray: AI on attack151,000+ failed logons across eight organizations. No successful sign-in. Lockouts made up nearly 98% of responses.
- 16–21 OctConnectWise Cloud Backup: AI on attackStolen app credentials, then machine-speed sign-ins to customer tenants. Microsoft attributes it to the China-state aligned Storm-0940.
- 21 Oct 2026OpenAI, Patch the Planet: AI on defenceAI found the critical PHP flaws before attackers did. OpenAI expects similar findings to “increase sharply in the coming months.”
- 21–22 OctQuest Diagnostics: no AI link reportedUnauthorized access to a lab ordering and results system; certain patient records accessed. No cause or attacker named.
Most AI-driven attempts in these reports failed. What changed is the cost of trying.Assessment: Thorsten Meyer
Sources: Unit 42 (30 Jul 2026); NSA, CISA, FBI, DOE, EPA advisory (18 Aug 2026); Microsoft Threat Intelligence (21 Oct 2026); ConnectWise Trust Center (from 16 Oct 2026); OpenAI, Patch the Planet (21 Oct 2026); Quest Diagnostics Form 8-K (22 Oct 2026).
Powered by Thorsten Meyer AI
Faster Attacks Raise Defensive Pressure
The significance of these reports is less that AI has produced a wholly new kind of attack than that it can help automate familiar tasks: reconnaissance, script development and repeated login attempts. Automation changes the time available to defenders. A cloud account or exposed industrial controller may require rapid detection and response even when attackers use broadly known techniques rather than a novel exploit.
That pressure applies to both security teams and ordinary users. Organizations need to account for trusted third-party applications, limit what those applications can access and review sign-in activity. Users can reduce the risk of account takeover by securing their email accounts with phishing-resistant sign-in methods where available, since email often helps control password resets and other services. The reports do not show that AI defeats these defenses; rather, they underscore the cost of leaving basic controls incomplete.
AI may also help defenders find vulnerabilities. OpenAI’s statement about flaws found before attackers is a company report, while its expectation of a sharp rise is a forecast, not a measured trend. If vulnerability discovery and disclosure accelerate, organizations could face more patches to evaluate and deploy. The available material does not quantify how many additional flaws will emerge or how quickly attackers will exploit them.
business firewall with intrusion prevention system
As an affiliate, we earn on qualifying purchases.
As an affiliate, we earn on qualifying purchases.
Three Reports, Different Evidence
The reports describe distinct kinds of evidence and should not be treated as one incident. The August government advisory concerns activity directed at Siemens PLCs and critical infrastructure. Microsoft’s October account concerns a cloud backup application and attempts to use its trusted access to reach customer tenants. A separate Unit 42 report dated July 30 described a Chinese-speaking actor using the DeepSeek model with the Hermes Agent framework, according to the source material; the details provided here do not establish a link between that case and the others.
The ConnectWise episode began with the company’s Oct. 16 disclosure, followed by Microsoft’s account on Oct. 21. Microsoft attributed the activity to Storm-0940 and described the role of ARTEX; those are the company’s assessments. ConnectWise’s statement about its archives is a separate company claim. Keeping those attributions clear matters: evidence of attempted access is not evidence of data theft, and claims about an actor’s identity remain assessments unless independently established.
The source material also cites a security leader predicting a “vulnerability tsunami” within months and estimating roughly 18 months to “hold on.” That is an expert’s opinion, not a measured forecast or a finding from the agencies’ advisory. The confirmed reports support concern about automation, but they do not establish a timetable for a broader surge.
“This is not a theoretical risk—it is an active threat.”
— Joint advisory by the NSA, CISA, FBI, Department of Energy and EPA
network security monitoring appliance for small business
As an affiliate, we earn on qualifying purchases.
As an affiliate, we earn on qualifying purchases.
What the Reports Do Not Establish
The reports do not establish that customer data was taken in the ConnectWise-related activity. Microsoft said a successful sign-in does not necessarily indicate data access, while ConnectWise said its existing backup archives showed no indication of compromise. The scope and outcome of the investigation were not fully detailed in the supplied material.
It is also unclear how often AI materially changes the outcome of attacks compared with conventional automation, or how many incidents across the wider threat landscape involve AI-generated code or agent frameworks. The government advisory describes activity against PLCs but does not, in the material provided, confirm operational damage. OpenAI’s expected increase in AI-discovered flaws and the security leader’s timeline for a surge remain forecasts, not established facts.
industrial control system cybersecurity firewall for PLC networks
As an affiliate, we earn on qualifying purchases.
As an affiliate, we earn on qualifying purchases.
Investigations and Basic Controls
ConnectWise’s investigation and Microsoft’s continuing threat analysis may clarify whether any tenant data was accessed and how the compromised app credentials were obtained. No further investigation milestone or final impact assessment is specified in the source material. Organizations using affected services should follow provider notices and review relevant access logs rather than treating attempted access as proof of a breach.
For industrial operators, the agencies recommend inventorying PLCs, applying firmware updates, keeping controllers off the public internet and separating operational technology from business IT. Their advisory also calls for multi-factor authentication on remote OT access and monitoring for controller write activity outside approved change windows. For companies more broadly, the practical priorities are asset inventories, timely patching, phishing-resistant sign-in, third-party app reviews and rehearsed incident response. Users can begin by protecting their primary email account and enabling strong authentication.
hardware security key for Microsoft 365 multi-factor authentication
As an affiliate, we earn on qualifying purchases.
As an affiliate, we earn on qualifying purchases.
Key Questions
Have AI-assisted attacks against industrial systems been reported?
Yes. A joint US agency advisory dated Aug. 18 says actors used AI-generated scripts against internet-reachable Siemens S7 controllers. It describes activity and a risk assessment, not confirmed operational damage in the material provided.
Was data stolen in the ConnectWise-related activity?
The available reports do not establish that. Microsoft said a successful sign-in does not necessarily mean data was accessed, and ConnectWise said its existing backup archives showed no indication of compromise.
Did the reported password-spray attempts succeed?
Microsoft said none of the more than 151,000 failed logons in the campaign from Oct. 10 to 14 succeeded. The company separately reported privileged cloud access in under three hours in several attacks, but the supplied account does not specify that this access resulted from those failed attempts.
What should organizations do now?
Follow the agencies’ guidance for industrial systems: inventory and patch controllers, prevent direct internet exposure, separate OT from IT and secure remote access. Across the organization, review third-party app permissions, strengthen authentication, monitor access and practise incident response.
Is a surge in AI-found vulnerabilities confirmed?
No. OpenAI said it expects the number to increase sharply in the coming months, but that is a forecast. The source material does not provide a measured increase or a confirmed timetable for a wider vulnerability surge.
Source: ThorstenMeyerAI.com
Evergreen bestsellers Picks
bestsellers
As an affiliate, we earn on qualifying purchases.
