The discovery of the human immunodeficiency virus (HIV) in breastmilk in 1985, and subsequent research, supports the hypothesis that breastfeeding provides a route of transmission to the nursing baby. Various routes of infection and relative rates of transmission have been studied in many parts of the world, leading to the blanket guideline that babies of HIV-infected mothers should not be breastfed, if a safe alternative can be provided. However, due to the limits inherent in various studies and various testing methods, the exact frequency of breastmilk transmission of HIV during the course of lactation remains unknown, and the conclusions drawn are thus conflicting and confusing. Replacement feeding of young babies with non-human milks and other foods may be hazardous in poverty-stricken populations in Africa and elsewhere, and still more research suggests that there are several properties in human milk that may provide specific protection to the baby of an infected mother. The possibility of providing the mother's own treated expressed breastmilk to the baby at risk of HIV infection via breastfeeding is an alternative which has yet to be fully explored and ways that this could be accomplished are examined. Those of us working with mothers and babies need more information before we can assist mothers living with HIV to make truly informed decisions about the safest way to feed their babies. Topics requiring urgent further attention are outlined.