Background: Gene therapies for articular cartilage defects are limited by the absence of an in vivo delivery system that can mediate site-specific transduction restricted to within the margins of the defect during routine arthroscopy. We have proposed the use of ultraviolet light to stimulate gene expression following infection by recombinant adeno-associated virus (rAAV). However, research has demonstrated that short-wavelength ultraviolet light (ultraviolet C), while effective, is neither safe nor practical for this purpose. We evaluated the safety and efficacy of long-wavelength ultraviolet light (ultraviolet A) from a laser to induce light-activated gene transduction in articular chondrocytes in vitro and in vivo. Methods: The effects of ultraviolet A from a 325-nm helium-cadmium laser, delivered through a fiberoptic cable, on cytotoxicity, mutagenesis, intracellular reactive oxygen species, and light-activated gene transduction of human articular chondrocytes were evaluated in dose-response experiments of primary cultures. Cytotoxicity was determined by trypan blue exclusion. The presence of pyrimidine dimers in purified genomic DNA was determined by enzyme-linked immunosorbent assays. Intracellular reactive oxygen species levels were determined by flow cytometry at one hour and twenty-four hours. In vitro light-activated gene transduction with rAAV vectors expressing the green fluorescent protein (eGFP) or β-galactosidase (LacZ) was determined by fluorescence microscopy and bioluminescence assays, respectively. In vivo light-activated gene transduction was quantified by stereotactic immunohistochemistry for β-galactosidase in rabbit articular cartilage defects in the patellar groove that had been irradiated with ±6000 J/m2 of ultraviolet A one week after direct injection of 107 transducing units of rAAV-eGFP. Results: Ultraviolet A failed to induce significant cytotoxicity at all fluencies below 6000 J/m2. Dose-dependent cytotoxicity was observed at greater fluencies. In contrast to ultraviolet C, which induced significant (p < 0.05) pyrimidine dimer formation at all fluencies in a dose-dependent manner, ultraviolet A failed to induce DNA modifications. Conversely, ultraviolet C proved to be a poor inducer of intracellular reactive oxygen species, while ultraviolet A immediately induced high levels of intracellular reactive oxygen species, which were completely resolved twenty-four hours later. Ultraviolet A demonstrated significant light-activated gene transduction effects in vitro, which were dose-dependent (p < 0.05). In vivo, ultraviolet A mediated a tenfold increase in transduction in which 40.8% of the superficial chondrocytes adjacent to the defect stained positive for green fluorescent protein compared with 5.2% in the knees treated with no ultraviolet A (p < 0.006). Conclusions: These results provide what we believe is the first formal demonstration of an agent that can induce rAAV transduction in the complete absence of cytotoxicity and DNA modification. They also suggest that the mechanism by which long-wavelength ultraviolet light mediates site-specific gene expression is by means of the induction of intracellular reactive oxygen species. Finally, laser-derived ultraviolet A can be readily transferred through a fiberoptic cable to mediate light-activated gene transduction in vivo. Clinical Relevance: This is the first demonstration of in vivo site-directed gene delivery to articular defects with use of a method that is highly compatible with standard arthroscopy. Future studies with chondrogenic genes and longer outcome measurements are warranted to evaluate the potential of this gene therapy approach for superficial articular defects and meniscal tears.