🔍 Read the full analysis: Cyberattacks Are Getting Faster With AI: Steps For Better Readiness 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 are using AI-generated scripts against internet-reachable Siemens industrial controllers, while Microsoft reports an AI agent was used in a campaign targeting cloud tenants through a compromised backup app. The reports document faster, more automated attack activity, but also say most attempts in the Microsoft cases failed; experts’ predictions of a coming “vulnerability tsunami” remain forecasts.
US agencies and Microsoft have reported separate cases of AI-assisted cyber activity, including scripts targeting industrial controllers and an agent used in an attack campaign against cloud tenants. The reports show attackers applying AI to familiar tasks such as reconnaissance, exploitation attempts and sign-in activity, while also recording that many attempts did not succeed.
A joint Cybersecurity Advisory published August 18 by the NSA, CISA, FBI, Department of Energy and Environmental Protection Agency says actors have used AI-generated scripts against internet-reachable Siemens S7 programmable logic controllers. These controllers manage industrial processes in sectors including manufacturing, energy, water, chemicals, food and agriculture. The agencies say the scripts were disguised as monitoring tools and built using the open-source snap7 library. They describe the activity as preparation that is likely intended to support operational effects, not as proof that physical damage has occurred.
In an October 21 report, Microsoft Threat Intelligence said the China-state-aligned actor it tracks as Storm-0940 used ARTEX, an open-source agentic penetration-testing framework, in activity tied to the ConnectWise Cloud Backup compromise. Microsoft says the actor used credentials associated with the cloud backup application to attempt sign-ins across more than 2,250 tenants. Microsoft also recorded more than 151,000 failed logons in a password-spraying campaign from October 10 to 14, peaking at 2,900 failures per minute; it said none of those attempts succeeded.
Microsoft said that in several attacks, the interval from initial targeting to privileged cloud access was under three hours. It also cautioned that most targeting was unsuccessful and that a successful sign-in or an affected application does not by itself show that data was accessed. ConnectWise said existing backup archives showed no indication of compromise and that it suspended Cloud Backup, SaaS Security and Cloud Management while investigating.
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 Response Pressure
The reports matter because they describe automation applied to ordinary attack steps, rather than establishing that AI has created a wholly new kind of cyberattack. Software can help attackers scan targets, adapt scripts or repeat sign-in attempts at a pace that can compress the time defenders have to identify activity and contain access. That pressure is especially relevant when attackers exploit trusted third-party applications or reach systems that control physical processes.
The operational lesson is not that every organization needs to buy an AI security product. The cited advisory and Microsoft findings point toward established controls: know what is exposed, limit remote access, separate industrial and corporate networks, review third-party app permissions and rehearse incident response. Faster attack workflows make delays in detection and patching more consequential, but they do not make prevention or response futile.
AI may also help defenders find flaws. OpenAI said it identified critical vulnerabilities in PHP that were patched on October 21, before attackers exploited them, and forecast that AI-discovered critical vulnerabilities will “increase sharply in the coming months.” That is a company forecast, not a measured prediction of how many flaws will be found or how quickly organizations will fix them.
Products Worth Considering
LG UltraGear 32GX870A-B 32-inch 4K UHD OLED Gaming Monitor
Samsung Odyssey G55C 32-Inch QHD 1000R Curved Gaming Monitor
LG UltraGear 27-inch QHD OLED Gaming Monitor, 240Hz
As an affiliate, we earn on qualifying purchases.
Two Documented Attack Pathways
The industrial-control advisory describes actors conducting reconnaissance against US installations of Siemens S7 controllers, from the S7-200 to the S7-1500, including safety controllers. According to the agencies, the systems can be found through internet-scanning services, and the activity includes AI-generated exploitation scripts. Their stated recommendations include inventorying controllers, updating firmware, keeping PLCs off the public internet, blocking TCP port 102 at the perimeter, separating IT and operational technology networks, and requiring multifactor authentication for remote OT access.
The Microsoft report concerns a different route: access through a trusted cloud application. Microsoft says ARTEX can conduct multiple stages of testing, from reconnaissance toward data exfiltration, with limited human interaction. The company’s report does not say that every attempted tenant access led to a breach. ConnectWise disclosed its compromise on October 16 and said it suspended the affected services during its investigation; Microsoft’s October 21 account linked Storm-0940 to the incident.
A separate July 30 report from Palo Alto Networks’ Unit 42 described 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 failed. Its report provides another example of AI being incorporated into an attack workflow, but the supplied account does not establish the full outcome of that activity.
“This is not a theoretical risk—it is an active threat.”
— NSA, CISA, FBI, Department of Energy and EPA joint advisory
Breach Impact Remains Unsettled
The available reports do not establish that the Siemens-related activity caused physical disruption or damage. The agencies assess that it may be preparation for operational effects, but that assessment is not confirmation that such effects occurred. The supplied material also does not specify the number of controllers compromised or the extent of any access.
For the ConnectWise incident, Microsoft says some attacks reached privileged cloud access in under three hours, but it cautions that access does not necessarily mean data was viewed or taken. ConnectWise said its existing backup archives showed no indication of compromise while it investigated. The full scope, any confirmed customer-data impact and the investigation’s final findings remain unclear in the material available here.
Predictions that the security industry faces a near-term “vulnerability tsunami,” or has roughly 18 months to prepare, are opinions attributed to an unnamed security leader. No measurement or methodology is supplied for those estimates. OpenAI’s forecast of more AI-found flaws is also forward-looking; its scale and implications for attackers and defenders remain unknown.
Investigations and Basic Controls
ConnectWise’s investigation will clarify whether the compromise affected customer data or backup archives and what further containment steps are needed. Organizations using the named services should follow vendor and Microsoft notices, review relevant sign-in and application activity, and avoid treating an attempted login as proof of a successful breach.
For industrial operators, the agencies’ advisory calls for action on exposed controllers, firmware, network separation and remote-access protections. For companies more broadly, readiness starts with an accurate inventory, defined patch deadlines, phishing-resistant sign-in where available, regular reviews of third-party application permissions, and incident-response exercises. These measures do not depend on predicting how capable future AI systems will be.
Individual users can begin by securing their email account, which often provides a route to reset passwords for other services. Use a unique password, turn on multifactor authentication, and review recovery details and active sessions. Further findings from the ConnectWise investigation and any updated government guidance may change the picture; the reports to date support preparation, not assumptions that every organization has been breached.
Key Questions
Have AI-assisted attacks caused confirmed physical damage to industrial systems?
The joint advisory reports AI-generated scripts targeting Siemens S7 controllers and says the activity may be preparation for operational effects. The supplied information does not confirm physical damage or disruption.
Did the ConnectWise activity expose customer data?
That remains unclear. Microsoft says some attacks reached privileged cloud access but cautions that access does not necessarily mean data was accessed. ConnectWise said existing backup archives showed no indication of compromise while it investigated.
What should industrial operators do first?
Follow the agencies’ recommendations to inventory controllers, update firmware, keep PLCs off the public internet, separate IT and operational technology networks, and protect remote access with multifactor authentication.
Does the reporting prove AI makes every cyberattack more successful?
No. The reports document AI being used in attack workflows, but Microsoft said most of the targeting in the cases it described was unsuccessful. The evidence supports concern about speed and scale, not a claim that AI guarantees success.
Is a coming “vulnerability tsunami” confirmed?
No. That phrase is an expert opinion reported by ThorstenMeyerAI.com. OpenAI separately forecast a sharp rise in AI-found critical vulnerabilities, but the forecast is not a measured count of future flaws.
Source: ThorstenMeyerAI.com
