Cyber Resilience Act
Reasonably foreseeable misuse
Legal definition Art. 3(25) CRA
“the use of a product with digital elements in a way that is not in accordance with its intended purpose, but which may result from reasonably foreseeable human behaviour or interaction with other systems”
Reasonably foreseeable misuse is the use of a product that runs against its intended purpose and that the manufacturer nonetheless has to expect. The Cyber Resilience Act (CRA) defines it in Article 3(25). Trace the term through the Regulation and you notice how sparingly it is used: outside the definition it appears exactly once, in Annex II, point 5.
The definition sets two elements. The first is the conflict with the intended purpose: the use is not in accordance with it. The second names the sources it may result from, namely reasonably foreseeable human behaviour or interaction with other systems.
How it differs from foreseeable use
Points 24 and 25 sit next to each other in the same article and are routinely mixed up. Three differences carry the distinction:
- Relationship to the intended purpose. Reasonably foreseeable use is “not necessarily” the intended purpose and may therefore coincide with it. Misuse is not in accordance with it.
- Threshold. Point 24 requires the use to be likely to result. Point 25 is satisfied where it may result.
- Sources. Point 24 names human behaviour, technical operations and interactions. Point 25 names human behaviour and interaction with other systems, but not technical operations.
The two therefore sit side by side rather than one inside the other: misuse applies the lower threshold, but it reaches only outside the intended purpose and leaves out technical operations. And it carries the narrower consequence. That is the scheme of the Regulation rather than an oversight: scope and risk assessment attach to foreseeable use, while misuse carries an information duty.
The one place it appears: Annex II, point 5
Annex II lists the information and instructions that must accompany the product. Point 5 requires any known or foreseeable circumstance, related to use in accordance with the intended purpose or under conditions of reasonably foreseeable misuse, which may lead to significant cybersecurity risks.
Two limits sit inside that sentence. What is asked for are circumstances, not behaviours: under which conditions something goes wrong, rather than a catalogue of conceivable operating errors. And the bar is a significant cybersecurity risk as defined in Article 3(38), a risk that, based on its technical characteristics, can be assumed to have a high likelihood of an incident that could lead to a severe negative impact. Anything below that bar does not belong in the list.
Article 13(18) governs the form: clear, understandable, intelligible and legible, in a language which can be easily understood by users and market surveillance authorities, and sufficient to allow secure installation, operation and use. The information has to stay available for at least 10 years after the product is placed on the market, or for the support period where that runs longer.
Through Annex VII, point 1(d), the information and instructions to the user are also part of the technical documentation. Whatever you wrote about misuse must therefore be kept at the disposal of the market surveillance authorities under Article 13(13) as well.
Interaction with other systems
The second branch of the definition is easy to overlook, because “misuse” sounds like a person pressing the wrong button. It equally covers another system addressing the product in a way its intended purpose does not provide for. Think of an interface built for a service tool and then polled by an automation every minute. Or a device wired into a plant whose network segment it was never meant for.
The German text repeats “reasonably foreseeable” before the interaction and so ties the qualifier to both sources. The English wording (“reasonably foreseeable human behaviour or interaction with other systems”) leaves open whether it reaches the interaction too. Little turns on it, since Annex II, point 5 asks only for known or foreseeable circumstances in any event.
What the CRA does not attach to misuse
Two provisions where the term would be expected do not name it:
- Scope. Article 2(1) turns on the intended purpose or the reasonably foreseeable use. A misuse that would create a data connection does not pull a product into scope.
- The risk assessment. Article 13(3) lists intended purpose and reasonably foreseeable use as minimum content. Misuse is not among them.
The second gap grants no relief. If Annex II, point 5 obliges you to state which conditions of misuse may lead to significant cybersecurity risks, those conditions must have been identified and assessed first. The cybersecurity risk assessment is where that work happens. The CRA simply does not require misuse to appear there under a heading of its own.
How far responsibility runs
Article 6(a) allows a product on the market only where it meets the requirements in Annex I, Part I, provided it is properly installed, maintained and used for its intended purpose or under conditions which can reasonably be foreseen and, where applicable, the necessary security updates have been installed. “Conditions which can reasonably be foreseen” is a third formulation, and the CRA does not define it. It reaches wider than either legal definition, because it speaks of conditions rather than of a use. Whether foreseeable misuse falls under it is not stated outright, though the wording points that way.
In practice the warning is the second answer, not the first. Annex I, Part I, point 2 requires, where applicable, a secure by default configuration with the ability to reset the product (point (b)), protection from unauthorised access (point (d)), and attack surfaces kept as small as possible, external interfaces included (point (j)). A misuse that a default setting can rule out does not become acceptable because the manual advises against it.
What stays open
When misuse counts as “reasonably” foreseeable, the Regulation does not say. There are no criteria, no examples, and the recitals do not develop the term. The line against attack stays open as well, since point 25 speaks of use and not of exploitation.
What makes the judgement defensible is evidence from the field: support cases about configurations nobody planned for, returns still on factory settings, deployments that sales knows about and engineering does not. Reviewing that material lets you justify the line you drew, and at this point the CRA asks for nothing more.
Practical questions
-
Only those that may lead to significant cybersecurity risks. Annex II, point 5 draws that line through Article 3(38): a risk that, based on its technical characteristics, can be assumed to have a high likelihood of an incident that could lead to a severe negative impact, including by causing considerable material or non-material loss or disruption. An exhaustive list of conceivable operating errors serves the provision worse rather than better, because Article 13(18) requires the information to be clear, understandable and legible. Three well-founded circumstances carry more than thirty precautionary ones.
-
The wording points the other way. Point 25 describes a use of the product, whereas an attack is the exploitation of a vulnerability by a cyber threat, which the CRA governs through its own set of provisions. The boundary is nowhere drawn expressly. The two do meet where misuse first creates the condition an attack needs, such as an administrative interface made reachable from the internet contrary to the intended purpose. Circumstances of that kind belong in the user information under Annex II, point 5, however you classify the attack itself.
-
In the first place the manufacturer, as far as the integration was foreseeable. Where the product is supplied as a component, the integrator is the user of your Annex II information, and the circumstances under which an installation leads to significant cybersecurity risks belong there. If the integrator carries out a substantial modification and makes the product available, Article 22 treats that person as a manufacturer, for the part affected or for the whole product where the change bears on its cybersecurity overall. Absent such a change, responsibility does not move.
This glossary is for orientation and does not constitute legal advice. The wording of Regulation (EU) 2024/2847 prevails.