Authentic vehicles can hide malicious functions despite protocol compliance
The claim
Authenticity does not guarantee declared-purpose conformance, allowing covert malicious capabilities to evade detection
Published as: “Janus Vehicles: Entity-Centric Threat Modelling of Protocol-Conformant Vehicles under a Malicious-Manufacturer Assumption”
So what
Automotive assurance must consider that a vehicle can be authentic and protocol-conformant while still violating its declared purpose, requiring new investigation and assurance approaches.
The five-minute version
The paper proposes an Entity-Centric Threat Model that defines four malign entity archetypes, a Behaviour-Locus Matrix, and a Behavioural Interaction Model to capture hidden, purpose-violating functionality in vehicles that remain protocol-conformant. It is a conceptual contribution only; the authors acknowledge no empirical evaluation or detection method is provided, so the evidence is purely theoretical.
What they found 3
- Authenticity does not guarantee declared-purpose conformance and malicious functionality can evade detection.
“The model exposes two assurance gaps: authenticity does not establish declared-purpose conformance, while malicious functionality can evade detection based on behavioural history or peer-visible evidence.”
- A Behaviour-Locus Matrix is introduced to classify actions across loci and roles.
“We introduce a Behaviour-Locus Matrix that crossclassifies covert, overt, and concealment actions across hardware, software, and network loci.”
- A Behavioural Interaction Model is provided to represent capability activation and contextual triggers for peer-benign, purpose-malicious vehicles.
“For peer-benign, purpose-malicious vehicles, we present a Behavioural Interaction Model that represents vehicle capabilities, contextual triggers, and their realisation through actions.”
What undercuts it the paper’s own results
- The model has not been empirically evaluated, limiting confidence in its practical applicability.
Stated limits 3
- This work presents a conceptual threat model rather than an operational investigation or detection method.
- The proposed model has also not yet been empirically evaluated using vehicles or testbeds containing seeded hidden functionality.
- The behaviours and manifestations presented in the Behaviour-Locus Matrix are illustrative rather than exhaustive, and a single capability may produce actions and evidence across several matrix cells.
What’s new here 3
- Defines four specific malign entity archetypes including malicious-by-design vehicles.
- Introduces a Behaviour-Locus Matrix that maps covert, overt, and concealment actions across hardware, software, and network loci.
- Presents a Behavioural Interaction Model linking capabilities, triggers, actions, artefacts, and concealment for purpose-malicious vehicles.
Artifacts & methods 3
- Malign Entity Archetypestheory
- Behaviour-Locus Matrixtheory
- Behavioural Interaction Modeltheory