Given a Vehicular Fog Computing (VFC) scenario, computing tasks with varying sensitivity levels must be offloaded while jointly balancing latency, energy, and trust requirements. Existing trust-based schemes often neglect some of these trade-offs or fail to adapt to task sensitivity and node reliability in dynamic vehicular settings. This paper proposes a trust-based task offloading (TTO) framework that employs a multi-dimensional trust model, combining Reputation, Experience, and Knowledge (REK) to compute a trust-aware overhead function that adapts to task requirements.We also incorporate computable encryption options, enabling task owners (TOs) to protect highly sensitive tasks. The interaction among TOs is modeled as a potential game, which implies the finite-improvement property and the existence of a Nash equilibrium, enabling distributed, self-adaptive decision-making that aligns individual incentives with overall system efficiency. This design allows vehicles to achieve an optimal balance between trust assurance and performance without centralized coordination. The efficiency of the framework against centralized approaches is validated by analyzing the price of anarchy. The simulation results confirm that TTO converges in 60 iterations for 100 nodes, achieving a 30-35% lower average delay and up to 25% lower total overhead compared to the considered baselines. These results indicate that TTO effectively balances trust, latency, and energy requirements.

We introduce the concept of an$\varepsilon $-semimetric that satisfies the same axioms as a topological metric, except for an arbitrarily small allowance to violate the triangle inequality. Under this modification, we demonstrate the possibility of taking arbitrary points in space, assigning arbitrary desired distances between them (independent of their geometric location relative to each other, that is, independent of their “features”), and constructing an$\varepsilon $-semimetric that measures exactly the desired distances in the point cloud. This results in a threat to fairness and objectiveness in applications of clustering algorithms: suppose that an adversary subjectively classifies people according to its whim or discriminatory preferences. Upon accusations of unethical behavior, the malicious data processor can plausibly deny these as follows: it designs a distance function (an$\varepsilon $-semimetric) that is (up to a fully controllable numeric “round-off-error”$\varepsilon $) equivalent to a standard distance like the Euclidean. However, this crafted distance will exactly reproduce the (malicious) results and thus confirm them while pretending objectivity and transparency, since only standard and explainable artificial intelligence was used. This demonstration works without any data poisoning. We illustrate the method on randomly chosen points with stochastically independent random classifications assigned to them. Then, we apply standard implementations of k-Means and DBSCAN on the data points, which both exactly reproduce the desired (randomly chosen) classes. We also discuss non-adversarial applications of$\varepsilon $-semimetrics, and corroborate the construction with examples and implementation in Octave.