Five agencies, one Siemens S7 warning, and a claim that AI is writing the exploit scripts

The reconnaissance in this advisory is not performed by the attacker. According to CISA, the actors use internet scanning services to find internet-exposed PLCs that are running outdated software or are otherwise poorly protected ^[5]. The quoted language describes a version-and-configuration condition rather than a single named flaw ^[5], which means the target list is generated continuously by third parties and refreshed without anyone attacking anything.

What historically sat between appearing in a scan result and being under someone else's control was labour: learning the protocol, matching firmware behaviour, building tooling that does not crash the device. The agencies say the actors are using publicly available information on these widely used controllers to develop exploits that permit remote access and control ^[1], and that AI tools are helping them identify additional attack vectors and possibly adapt to defensive measures operators have already put in place ^[9]. Read as an engineering statement, that is a claim about cost. The scanning half of the kill chain was already cheap; the advisory says the bespoke half is getting cheaper too.

The detail that undercuts the mitigation list is further down. Agencies say AI tooling can make malicious files look and behave like legitimate monitoring tools, built on open-source industrial automation libraries ^[10]. The fourth item operators are told to do is deploy cybersecurity measures that monitor industrial control systems for anomalies and possible malicious activity ^[4]. An anomaly detector tuned to spot malformed or unusual S7 traffic is being asked to flag a well-formed client using the same libraries an integrator would use. Of the four recommended controls, patching, isolation, access control and monitoring ^[4], the one that still works cleanly against the described adversary is the one that removes the device from the scan results in the first place.

The advisory also puts the loss in two ledgers at once: disruption of critical industrial processes, safety incidents, downtime or equipment damage, compromise of sensitive data, compliance violations, and cascading impacts across interconnected systems ^[7]. CISA states plainly that this is not a theoretical risk but an active threat ^[3].

On attribution, the reporting is not internally consistent. Tom's Hardware describes the warning as concerning Iranian hackers, then states the agencies did not specify where the attacks could originate ^[11]. It also notes the advisory landed less than a month after water infrastructure in several states was hit by cyberattacks thought to have come from Iran ^[12]. For an operator with an S7 answering on a public address, the flag on the origin changes nothing about the work.