Why AI Has Changed The Timing Of Cyberattacks For Companies And Users

🔍 Read the full analysis: Why AI Has Changed The Timing Of Cyberattacks For Companies And Users 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

Recent government advisories and threat reports describe attackers using AI-generated scripts and agentic tools to target industrial controllers and cloud services, raising concern that attacks can move faster and operate for longer. The reports document specific activity, but do not show that AI makes attacks automatically successful: Microsoft said most of the targeting it tracked failed, while OpenAI separately forecast a sharp rise in vulnerabilities found with AI.

In an August 18 joint advisory, the NSA, CISA, FBI, Department of Energy and Environmental Protection Agency said actors were using AI-generated exploitation scripts against internet-reachable Siemens S7 programmable logic controllers. The agencies described reconnaissance and capability development affecting critical infrastructure sectors, including manufacturing, energy, water and food and agriculture. They assessed the activity as likely preparation for possible operational effects; that assessment is not confirmation that such effects occurred.

Microsoft Threat Intelligence reported on October 21 that Storm-0940, which it describes as a China-state-aligned actor active since 2021, used the open-source ARTEX multi-agent framework in activity associated with the ConnectWise Cloud Backup compromise. Microsoft said the actor tried to sign in to more than 2,250 customer tenants using credentials for the cloud backup app. From October 10 to 14, Microsoft recorded more than 151,000 failed logons, peaking at 2,900 per minute; none of those attempts succeeded. In several attacks, Microsoft said, the interval from initial targeting to privileged cloud access was under three hours.

Microsoft cautioned that most targeting attempts were unsuccessful and that an affected app or successful sign-in does not by itself show that data was accessed. ConnectWise, which disclosed the compromise on October 16, said existing backup archives showed no indication of compromise. The company suspended Cloud Backup, SaaS Security and Cloud Management while it investigated. Separately, OpenAI said AI had helped identify critical PHP vulnerabilities that were patched on October 21, and forecast that AI-found vulnerabilities would increase sharply in coming months. That is a forecast, not a measured count of future vulnerabilities.

Faster Workflows Raise the Response Stakes

The reports point to a change in the time available to defenders: software can automate reconnaissance, repeated login attempts and parts of exploit development without the pauses that constrain human operators. Microsoft’s account of privileged access in under three hours illustrates why organizations cannot rely on slow, manual review of alerts or wait for a routine patch cycle when a serious exposure is identified.

That does not mean AI has made every attacker more capable or attacks more successful. In Microsoft’s reported campaign, the large volume of failed logons did not produce a successful result. The practical concern is that automation may lower the effort needed to repeat ordinary attack steps across many targets, while also increasing the pace at which vulnerabilities are discovered and fixes become necessary. Companies may face more pressure to maintain accurate inventories, prioritize urgent patches and rehearse incident response.

For individuals, the same reporting reinforces the value of protecting accounts that can reset or grant access to others. Email accounts are often central to password resets and notifications. Unique passwords, a password manager and phishing-resistant multi-factor authentication where available can reduce exposure, though they cannot prevent every compromise.

Three Reports Show Different Uses

The source material describes several separate cases, not one coordinated campaign. The August advisory concerns industrial control systems, including Siemens S7 controllers. The agencies said actors found exposed systems through internet scanning services and used AI-generated scripts disguised as monitoring tools, built with the open-source snap7 library. Their recommendations include keeping controllers off the public internet, updating firmware, separating operational technology from corporate IT, requiring multi-factor authentication for remote access and monitoring for unauthorized control-system writes.

Microsoft’s October report concerns a different pathway: abuse of a trusted cloud application and its credentials. ConnectWise’s response and Microsoft’s caveats matter because the report distinguishes attempted access from confirmed data theft. A separate July 30 report from Palo Alto Networks’ Unit 42 described an actor using the DeepSeek model with the Hermes Agent framework, controlled through Telegram; the supplied material says the agent searched GitHub after an initial exploit failed. Unit 42 reported that one agent performed hundreds of hours of manual targeting analysis in minutes. That is the company’s observation, not evidence that all AI agents perform similarly.

Across these accounts, the common thread is automation of familiar tasks rather than an independently operating “AI super-hacker.” AI can assist with scanning, adapting tools or analyzing targets, while human operators and existing access paths remain part of the reported activity. On the defensive side, OpenAI’s account of AI-supported vulnerability discovery shows that the same broad technology can help find flaws before attackers exploit them.

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

— NSA, CISA, FBI, Department of Energy and EPA joint advisory

Success and Future Scale Remain Unclear

The reports do not establish how often AI-assisted activity leads to a successful intrusion, how much of the reported work could have been done with conventional automation, or whether AI was decisive in each operation. Microsoft said most targeting failed and that the ConnectWise-related activity did not, on its own, confirm data access. ConnectWise’s statement about its existing backup archives is the company’s account; the supplied material does not provide a final investigation finding.

The agencies’ assessment that activity against controllers was likely preparation for operational effects describes their judgment about intent, not a confirmed disruption. Nor is there a reliable overall measure here of how frequently attackers use AI. OpenAI’s expectation of a sharp increase in AI-discovered vulnerabilities and a security leader’s warning of a coming “vulnerability tsunami” are forecasts. The latter is an expert opinion, not a measured trend or established timeline.

Patch, Contain and Verify Access

Organizations should follow the relevant advisories, verify whether exposed controllers or affected cloud applications are present in their environments, and check for unusual access or configuration changes. For operational technology, the agencies’ guidance calls for an accurate controller inventory, firmware updates, restricted internet exposure, separation between IT and OT, multi-factor authentication on remote access and monitoring for unexpected control traffic.

For cloud services, security teams should review connected applications and granted permissions, investigate unusual sign-ins, and follow vendor updates as incident investigations proceed. Boards and executives can ask for evidence of patch deadlines, tested backups, phishing-resistant sign-in and rehearsed incident response rather than treating AI adoption itself as a security plan. Users can start by securing their primary email account and reviewing recovery methods.

The next meaningful developments are final findings from the ConnectWise investigation and further technical details from agencies and security researchers. As AI-assisted vulnerability discovery produces new findings, organizations will need to judge each flaw by its exposure and severity rather than assume every disclosure signals an active attack.

Key Questions

Do the reports show that AI cyberattacks are usually successful?

No. Microsoft said most targeting in the activity it described was unsuccessful; it also recorded more than 151,000 failed logons, none of which succeeded. The reports document AI use, not a general success rate for AI-assisted attacks.

What systems did the US agencies warn about?

The August joint advisory described AI-generated scripts targeting internet-reachable Siemens S7 programmable logic controllers, used in industrial and critical-infrastructure settings. The agencies recommended restricting exposure and strengthening OT security.

Was customer data confirmed stolen in the ConnectWise incident?

The supplied reporting does not confirm that customer data was accessed. Microsoft said an affected app or successful sign-in does not necessarily mean data was accessed, and ConnectWise said its existing backup archives showed no indication of compromise while its investigation continued.

What should companies do now?

Organizations should inventory exposed systems, prioritize and apply relevant patches, review third-party app access, protect remote access with strong authentication, separate IT and OT networks where appropriate, and practise incident response. These measures address the reported risks without relying on predictions about future AI capabilities.

Does AI also help defenders?

Yes. OpenAI said AI helped identify critical PHP vulnerabilities patched on October 21. Its prediction that such findings will increase sharply is a forecast, but the reported discovery illustrates that AI can support defensive vulnerability research as well as attacker workflows.

Source: ThorstenMeyerAI.com