RNA interference describes the recently discovered process of sequence-specific, post-transcriptional gene silencing that is initiated by double-stranded RNA molecules known as small interfering RNAs (siRNAs). siRNAs have an acceptable half-life in vitro, a predictable biodistribution profile similar to that of single-stranded antisense oligonucleotides (ASOs), and have repeatedly been more robust than ASO techniques in terms of consistency of transcript knockdown and threshold concentration. Following validation in mammalian cells by Tuschl and co-workers in 2001, synthetic siRNAs have gained wide acceptance as a laboratory tool for target validation. Currently, there is considerable interest in the therapeutic use of siRNA, particularly in areas of infectious disease and cancer. In vitro and in vivo findings demonstrate the efficacy of siRNA knockdown of gene messages that are pivotal for tumor cell growth, metastasis, angiogenesis and chemoresistance, leading to tumor growth suppression. However, siRNA-based cancer therapy faces similar pharmacokinetic limitations to ASO therapy with respect to the extent that siRNA accesses primary and metastatic target cells. The recently identified 'off-target activity' of siRNAs is also of concern. The concept of carrier-restricted delivery of siRNA by conditionally replicative, oncolytic adenoviruses is discussed. Oncolytic adenoviral delivery offers the potential benefits of restricted and renewable siRNA expression within the tumor microenvironment, an additive antitumor outcome through viral oncolysis and siRNA-mediated oncogene silencing, and a proven clinical platform with respect to infectivity and safety.