I think it depends on what your conceptualization of the semantic network is. Suppose the final expression of the semantic network is to make heterogeneous sources of information available through web-like publishing mechanisms to allow agent programs to consume them in order to satisfy some high-level user target in an autonomous manner. This is close to Berners-Lee et al original conceptualization of the goal of the semantic network. In this case, agents should be aware that the information they receive from the three RDF repositories, SPARQL endpoints, rule bases, etc., is reliable, accurate, and trustworthy. The semantic network stack claims that a necessary step to reach this endpoint is to have a logic or collection of logics that the agent can use when it talks about acquired knowledge. It is rather a strong AI , or good to this end of the spectrum.
However, there is an alternative conceptualization (and in fact, probably a lot), in which the upper layers of the semantic web stack, including unifying logic, are not needed, because this is not what we ask the agents to do. In this view, the semantic network is a way to publish disaggregated, informative information for consumption by programs, but not autonomously. These are developers and / or users who choose, for example, which information is treated as trustworthy. This is related data , and it follows that the current stack of standards and technologies is ideal for creating useful applications. Indeed, some argue that even generally accepted standards, such as OWL , are not required to create data-bound applications, although I personally find it necessary.
Regarding responsibility, if you take the old point of view, this is something that the software agent community is already working on, and if you take the last opinion, it doesn’t matter if something is ever standardized, because we can start building useful functionality without him.