Richard Baxter famously stated that “the first line of defence against international humanitarian law is to deny that it applies at all”. While “under-classification” remains an issue today, a parallel trend needs to be acknowledged. This is the tendency to over-classify situations of violence, especially in relation to transnational terrorist organizations such as the so-called Islamic State group or Al-Qaeda. This tendency stems from practical difficulties inherent in the changing operational environment. The last few years have witnessed a proliferation of armed non-State actors that are labelled or designated as terrorists (e.g., in Iraq, Syria, Mali, Nigeria and Yemen). Terrorist groups are characterized by opaque, often volatile organizational structures and tend to operate in decentralized networks rather than clear hierarchies. The formation of splinter groups, changing alliances, temporary reunification and even open hostility among former allies are common phenomena. This complex factual situation has led to the proliferation of theories of conflict classification, many of them arguing in favour of more flexible classification via the loosening of existing standards. Because international humanitarian law is in many respects less protective than international human rights law, particularly regarding the rules on the use of force and detention, classifying a situation of violence as an armed conflict when the threshold has not been met is a problem that should not be underestimated. In this article, we revisit the criteria of intensity and organization, as well as the related matter of the role of motives in conflict classification, considering conflicts involving armed groups described as terrorists. Our goal is to identify minimum requirements that could diminish the risk of over-classification by various stakeholders.