Well, I'd have to read the study in order to comment intelligently on this particular case, but yes, gene therapies often rely on a tamed version of a highly infectious virus. Adenoviruses have long been used, though with substantially more caution now after the death of Jesse Gelsinger in 1999 from a massive immune response triggered by the dangerously high adenoviral epitope concentrations used in gene therapy on him. HIV does have the capacity to enter cell types that certain other viruses can't, and it is naturally a retrovirus (so it carries its own reverse transcriptase and can integrate into the host chromosome with relative ease), both of which make it a good candidate for a vector under some circumstances. I would hope, of course, that several of the dangerous genes the virus normally carries were fully removed (rather than inactivated by small mutations), and quite possibly that the extremely high realized mutation rate in the virus was dealt with.
Also of note is that the patient be extremely careful about potential HIV exposure. He should test positive for HIV for the rest of his life, given that a) routine HIV tests check for presence of antibodies to the virus, which he'll have, and b) even if they use a screen for the virus itself, they'll most likely screen for a protein or a gene that the gene-therapy-vector HIV makes/carries, and that virus will be living in the patient most likely until his death. As such, he won't have the early warning of the viral infection that many get these days, and will be much more vulnerable to a rapid development of AIDS if he does contract the untamed virus.