In 1975, when genetic engineering was still young, the leaders in the field called a meeting at Asilomar, a seaside conference centre in California, where they thrashed out the possible environmental and health risks of the powerful new gene-splicing techniques that they were wielding. They not only agreed important containment guidelines for certain kinds of work, but achieved something potentially more valuable: the wide press coverage they received won the public's trust that scientists were behaving responsibly.

Today that trust is on shaky ground. Controversies over genetically engineered crops and embryo research are leading people to question how carefully scientists consider the possible consequences of their work before barrelling ahead. This is no small concern for science, as it has already led to restrictions.

At the same time, biologists have come to feel increasingly secure in the belief that some ecological nightmare is not likely to spring out of a graduate student's Petri dish. Every day for decades they have been transferring modified genes into microbes, nematodes and mice. At least some of the results — the errant fruitfly or the culture tube spilled in the sink — have no doubt escaped into the environment, without producing a biological Chernobyl.

Is that confidence in step with the technology? The tools now available to the molecular biologist have the potential to provide a stunning array of benefits, for both biomedicine and basic biology. Researchers are learning to understand and manipulate the genetic circuits that control cells. They can transfer entire synthetic pathways to bacteria to make drugs that must otherwise be extracted from rare plants at great cost. Viral genomes can be synthesized chemically in weeks, and bacterial genomes will soon be within reach.

Through such technologies, a new field of synthetic biology is emerging (see page 624). Bacteria and yeast have been engineered to build proteins impossible in nature, and with novel properties, by the addition of synthetic amino acids. Several groups are even working on assembling simple cells from basic components. This is no longer a matter just of moving genes around. This is shaping life like clay.

Members of the synthetic-biology community have begun to discuss the possible risks, and ethical implications, of their work. But there is no plan as yet for anything like another Asilomar. In one sense, it may be too soon. The scope of these tools is much broader than that of recombinant DNA, and it is certain to be more difficult to foresee what the actual risks are.

But perhaps such discussions can't come soon enough. What will happen if biologists announce that they have made the first living cells from scratch without having demonstrated to the public any concern for the implications? Researchers must do more than talk among themselves. They must demonstrate publicly that they are willing to consult and reflect carefully about risk — both perceived and genuine — and to moderate their actions accordingly. The need for public trust, significant in 1975, is all the greater today.