The ability to detect ambiguities in context-free grammars is vital for their use in several fields, but the problem is undecidable in the general case. We present a safe, conservative approach, where the approximations cannot result in overlooked ambiguous cases. We analyze the complexity of the verification, and provide formal comparisons with several other ambiguity detection methods.