NSA, CISA, the FBI, the Department of Energy and the EPA issued a joint advisory this week warning that threat actors are using AI to generate exploitation scripts against internet-exposed Siemens S7 Series programmable logic controllers. Advisory AA26-231A describes the activity as “an active threat,” not a theoretical one, and names critical manufacturing, energy, water and wastewater, chemical, food and agriculture, and commercial facilities as the sectors being targeted. It is one of the first times a US agency has formally attributed an operational OT campaign to AI-assisted tooling rather than manually written exploit code.
What the advisory describes
According to the joint alert, attackers are using AI assistance to build Python-based exploitation scripts around the open-source snap7.dll and python-snap7 libraries, disguising the tools as legitimate operational technology monitoring software. The scripts communicate with targeted controllers over the S7comm protocol, giving an attacker who reaches an exposed device read and write access to PLC memory, configuration data, and ladder logic programs. Reconnaissance reportedly leans on internet-scanning services such as Censys and ZoomEye to locate exposed S7 devices before deploying the scripts. Affected models span the full S7 product line still in service across industrial environments: S7-200, S7-300, S7-400, S7-1200 (CPU 1211C through 1217C), and S7-1500, including F-series safety controllers.
The advisory’s own framing is notable: it states that AI “enables adversaries to rapidly leverage additional attack vectors and adapt to defensive measures,” lowering the technical bar for building functional ICS exploitation tooling and cutting development time. Siemens has pushed back on the novelty angle, telling reporters it has not identified new vulnerabilities or increased attack volumes in its ICS products, and an ISAC sector director quoted by Cybersecurity Dive attributed most of the reported activity to weak security processes and hard-to-maintain legacy hardware rather than a new exploit class. That distinction matters for executives: the risk here is less a zero-day and more that AI tooling is making it faster and cheaper to weaponize known misconfigurations and exposed protocols that have existed for years.
What it means for OT/ICS and critical infrastructure
The advisory arrives against a backdrop of sustained PLC-focused campaigns against US water, energy and manufacturing operators through 2026, including the Minnesota water utility intrusions and CISA’s earlier expansion of an Iranian-linked PLC advisory to cover Rockwell Automation, Schneider Electric, and Siemens equipment. This warning broadens that pattern by focusing specifically on AI-assisted tooling rather than a single threat actor, and by naming Siemens’ S7 line — among the most widely deployed PLC families in the world — as the current target set. Any organization running S7 controllers reachable from the internet, whether directly or through poorly segmented remote-access paths, should treat exposure as an active risk rather than a compliance gap to close later.
Recommended actions
Inventory every Siemens S7 device on the network and confirm none are directly internet-facing. Apply the latest Siemens security updates and firmware to all identified controllers. Enforce strict access controls and network segmentation between IT and OT environments, and remove any PLC currently reachable from the public internet immediately. Deploy monitoring for anomalous S7comm traffic, unexpected read/write activity against ladder logic, and connections resembling legitimate monitoring tools that were not deployed by the organization. Hunt retroactively for indicators consistent with the scripts described in AA26-231A, particularly unauthorized use of snap7-based libraries on engineering workstations or jump hosts with OT network access.