Fertility specialists derive enormous satisfaction from their work. Whereas many doctors are confronted daily with patients who are struggling against death, the fertility expert's stock-in-trade is the creation of life. It must be uplifting to arrive at work each morning and glance up at a wall filled with smiling family portraits that are the direct result of your professional endeavours.

But this emotional reward, along with the commercial pressures that bear on private reproductive clinics, means that caution is not always the watchword. Many clinics tout for business by quoting their success rate, so they are eager to adopt techniques that boost the chance of a successful pregnancy. At the same time, a desire to help more couples experience the joys of parenthood has led reproductive specialists to adopt more aggressive techniques to treat infertility. Intracytoplasmic sperm injection (ICSI), for instance, in which a sperm cell is injected directly into the egg, is now a routine procedure.

Some researchers are now starting to question the safety of ICSI and other techniques, claiming that they are linked with increased rates of birth defects and rare 'genetic imprinting' disorders (see page 656). This sounds alarming, but it's important to keep the fears in perspective. So far, the few studies done are mostly based on small sample sizes, and in some cases the findings are contested. Other studies on children conceived by assisted reproduction have found no evidence of any problems, and some developmental biologists see no reason to suppose that such techniques should pose a significant risk.

But there are good reasons to urge caution. We know from studies of livestock, where the manipulation of eggs and embryos is often more severe than in human fertility clinics, that such interventions may be associated with a syndrome in which fetuses grow too large, and may die at around the time of birth. Disturbingly, there seem to be parallels between this condition and the imprinting disorders now being linked tentatively to assisted human conception.

The tendency of human fertility specialists to push the boundaries ever outwards is a further cause for concern. Consider, for example, the sorry tale of cytoplasmic transfer, pioneered by Jacques Cohen of the Institute for Reproductive Medicine and Science of Saint Barnabas in Livingston, New Jersey. By injecting their eggs with cytoplasm from eggs donated by younger women, Cohen enabled some infertile women to have a child. But the technique also seems to heighten the risk of a chromosomal abnormality.

Few would argue that such adventures are desirable, but how can they be curbed? Requiring all new assisted-reproduction techniques to undergo full clinical trials would probably bring progress to an end. But Nature has argued previously that there is much to commend the British model, under which a statutory authority regulates fertility clinics and researchers (see Nature 420, 1; 200210.1038/420001a). In many countries, clinics in the private sector are given too free a rein.

There is also a need for more research to assess the risks posed by assisted reproduction. This will require further epidemiological follow-up and studies to determine whether the embryos created bear any cellular or molecular abnormalities. The current US administration, unfortunately, is unlikely to provide federal dollars for research on human embryos. For the sake of future generations of assisted-reproduction children, funding bodies in more permissive countries should rise to this important challenge.