How Businesses And Users Can Adapt To AI-Driven Cyber Risks

🔍 Read the full analysis: How Businesses And Users Can Adapt To AI-Driven Cyber Risks 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 reported AI-assisted cyber activity targeting industrial control systems and cloud accounts, including scripts aimed at Siemens controllers and an agent framework used in attacks involving a compromised backup app. The reports matter because they describe automation applied to familiar attack methods, while also stressing that many attempts failed and that a successful sign-in does not by itself prove data was accessed.

A joint advisory published August 18 by the NSA, CISA, FBI, Department of Energy and EPA says attackers used AI-generated exploitation scripts against internet-accessible Siemens S7 programmable logic controllers. PLCs are industrial computers used to control equipment and processes. The agencies say the scripts were disguised as monitoring tools and built using the open-source snap7 library. They assess that the activity may be preparation to cause operational effects against US critical infrastructure.

On October 21, Microsoft Threat Intelligence described activity by Storm-0940, which it identifies as a China-state-aligned actor active since 2021. Microsoft said the actor used ARTEX, an open-source multi-agent penetration-testing framework, in activity it linked to the ConnectWise Cloud Backup compromise. After obtaining credentials for the app, the actor attempted sign-ins across more than 2,250 tenants. Microsoft recorded more than 151,000 failed logons during a password-spraying campaign from October 10 to 14, with a peak of 2,900 failures per minute; it said none succeeded.

Microsoft said some attacks reached privileged cloud access in under three hours, but cautioned that most targeting was unsuccessful. It also said an affected app or successful sign-in does not necessarily mean 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 Raises the Pace of Attacks

The reported cases suggest AI can reduce the time and effort needed for routine parts of an intrusion, such as reconnaissance, password spraying and adapting exploitation attempts. The change described is operational speed, not evidence of an all-powerful or autonomous attacker. That speed can compress the time defenders have to spot unusual activity, verify an alert and contain access.

Industrial systems carry particular risks because unauthorized changes can affect physical operations, including manufacturing, energy, water and other infrastructure. Cloud attacks can spread through trusted third-party access: customers may have granted an application permissions before its credentials were compromised. The Microsoft figures also show why scale needs careful interpretation: a large number of failed attempts is not a count of successful breaches.

The defensive recommendations in the agency advisory are established controls rather than AI-specific products: keep an accurate inventory, patch firmware, prevent direct internet access to controllers, separate operational technology from business networks, require multifactor authentication for remote access and monitor control-system traffic. For organizations, basic security discipline remains the immediate response.

Two Documented Attack Pathways

The industrial-control advisory names Siemens S7 models ranging from the S7-200 to the S7-1500, including safety controllers. It says actors used internet-scanning services to locate exposed devices. The agencies recommend blocking TCP port 102 at network perimeters, using a demilitarized zone between IT and operational technology, and considering unidirectional gateways for historian connections where appropriate. They also recommend watching for S7 write commands outside approved maintenance windows.

The cloud case involved a different route: credentials associated with a trusted backup application and attempts to use that access across customer tenants. ConnectWise disclosed the compromise on October 16, according to the source material. Microsoft’s October 21 report attributed the activity to Storm-0940 and described ARTEX as capable of supporting stages from reconnaissance to data exfiltration with limited human interaction. The report does not establish that data was exfiltrated in these attempts.

Separately, OpenAI said AI had identified critical PHP vulnerabilities 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 from the company, not a measured prediction of how many flaws will be found or exploited.

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

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

What the Reports Do Not Establish

The public accounts do not establish that the industrial activity caused physical disruption, or that every targeted controller was compromised. The agency advisory describes preparation and capability development and assesses that actors may be preparing for operational effects. That assessment is not confirmation that such effects occurred.

For the ConnectWise-related activity, Microsoft says most targeting failed and that a successful sign-in does not necessarily indicate data access. ConnectWise reported no indication of compromise in existing backup archives, but the investigation status and the full scope of any customer impact are not specified in the supplied material. Actual data access and exfiltration remain unconfirmed in these accounts.

A security leader quoted in the source material predicted a “vulnerability tsunami” within months and said organizations might have roughly 18 months to prepare. This is an expert opinion, not a measured trend or an official timeline. The scale and pace of future AI-enabled attacks remain uncertain.

Practical Steps While Reviews Continue

Organizations should review whether industrial controllers are exposed to the internet, update firmware, restrict remote access and separate operational networks from corporate IT. They should check third-party applications for excessive permissions, review credentials and sign-in logs, and practice incident-response procedures. The advisory also calls for multifactor authentication on remote OT access and monitoring for unauthorized control-system writes.

For individual users, a practical starting point is to protect the email account, which often serves as a recovery channel for other services. Use a unique password, turn on multifactor authentication—preferably a phishing-resistant method where available—and scrutinize unexpected login approvals or messages requesting credentials. Businesses should set clear patch deadlines and confirm that staff know how to report suspicious activity.

ConnectWise’s investigation and any further agency or Microsoft updates may clarify the scope of the cloud incident. OpenAI’s vulnerability forecast will be tested over time by reported discoveries and patch activity; the supplied information does not give a timetable or numerical estimate. For now, organizations and users can act on the documented risks without treating forecasts as confirmed outcomes.

Key Questions

Have AI-assisted attacks caused confirmed physical damage to industrial systems?

The supplied advisory describes AI-generated scripts targeting Siemens controllers and assesses that activity may be preparation for operational effects. It does not confirm physical damage or disruption.

Did the ConnectWise-related activity expose customer data?

The available accounts do not confirm data access or exfiltration. Microsoft said a successful sign-in does not necessarily indicate data was accessed, and ConnectWise said its existing backup archives showed no indication of compromise.

What should a business do first?

Start by inventorying exposed systems and third-party applications, patching firmware, restricting internet and remote access to operational technology, enabling multifactor authentication and reviewing sign-in activity.

What can individual users do to reduce risk?

Secure email with a unique password and multifactor authentication, use phishing-resistant sign-in where available, and avoid approving unexpected login requests or sharing credentials through unsolicited links.

Is a surge in AI-discovered vulnerabilities confirmed?

No. OpenAI said it expects AI-found critical vulnerabilities to increase sharply in coming months. That is a forecast, not a confirmed count or a guarantee that every discovered flaw will be exploited.

Source: ThorstenMeyerAI.com