Recent reports have demonstrated that the group I ribozyme from Tetrahymena thermophila can perform trans-splicing reactions to repair mutant RNAs. For therapeutic use, such ribozymes must function efficiently when transcribed from genes delivered to human cells, yet it is unclear how group I splicing reactions are influenced by intracellular expression of the ribozyme. Here we evaluate the self-splicing efficiency of group I introns from transcripts expressed by RNA polymerase II in human cells to directly measure ribozyme catalysis in a therapeutically relevant setting. Intron-containing expression cassettes were transfected into a human cell line, and RNA transcripts were analyzed for intron removal. The percentage of transcripts that underwent self-splicing ranged from 0 to 50%, depending on the construct being tested. Thus, self-splicing activity is supported in the mammalian cellular environment. However, we find that the extent of self-splicing is greatly influenced by sequences flanking the intron and presumably reflects differences in the intron's ability to fold into an active conformation inside the cell. In support of this hypothesis, we show that the ability of the intron to fold and self-splice from cellular transcripts in vitro correlates well with the catalytic efficiency observed from the same transcripts expressed inside cells. These results underscore the importance of evaluating the impact of sequence context on the activity of therapeutic group I ribozymes. The self-splicing system that we describe should facilitate these efforts as well as aid in efforts at enhancing in vivo ribozyme activity for various applications of RNA repair.