Since testing for HIV infection became possible in 1985, testing of pregnant women has been conducted primarily on a voluntary, 'opt-in' basis. Faden, Geller and Powers, Bayer, Wilfert, and McKenna, among others, have suggested that with the development of more reliable testing and more effective therapy to reduce maternal-fetal transmission, testing should become either routine with 'opt-out' provisions or mandatory. We ask, in the light of the new rapid tests for HIV, such as OraQuick, and the development of antiretroviral treatment that can reduce maternal-fetal transmission rates to <2%, whether that time is now. Illustrating our argument with cases from the United States (US), Kenya, Peru, and an undocumented Mexican worker in the US, we show that when testing is accompanied by assured multi-drug therapy for the mother, the argument for opt-out or mandatory testing for HIV in pregnancy is strong, but that it is problematic where testing is accompanied by adverse events such as spousal abuse or by inadequate intrapartum or follow-up treatment. The difference is not a 'double standard', but reflects the presence of conflicts between the health interests of the mother and the fetus--conflicts that would be abrogated by the assurance of adequate, continuing multi-drug therapy. In light of these conflicts, where they still occur, careful processes of informed consent are appropriate, rather than opt-out or mandatory testing.