For scientists working with animals, human data or stem cells, or collecting ecological data in resource-poor settings, the ethical considerations of their research are something they must grapple with regularly. For those working at the molecular or cellular level or purely in silico, however, their experiences with ethics may be limited to a brief training course mainly covering issues of data manipulation and plagiarism.

Many molecular, cellular and computational biologists may think that research ethics — that is, whether the scientific questions being explored are ethical to pursue, whether the approaches used to pursue such questions are ethical, and whether the communication of that research is ethical — just do not apply to their work. But in this issue of Nature Methods, a Comment from Jeantine Lunshof and Julia Rijssenbeek1 implores life scientists of all stripes to integrate discussions with professional ethicists and philosophers into the research planning process.

For some research fields, ethical guidelines are readily available. Research on human subjects — for example, the use of fMRI to study alterations to brain connectivity networks in disease — must be approved by ethics committees (in accordance with the Declaration of Helsinki), and such subjects must provide informed consent. Experiments with lab mice must also be preapproved by an ethics committee, and researchers should follow the ARRIVE reporting guidelines2, as well as the American Veterinary Medical Association guidelines for animal euthanasia. Developmental biologists working with human embryos should abide by ethics guidelines set by the International Society for Stem Cell Research. Ecologists collecting data in the global South are encouraged to follow the TRUST Code recommendations — in particular, by including local researchers in the work.

There are many areas of life sciences research, however, where research does not require ethics oversight and ethics guidelines are not readily available, but where ethical considerations may yet be important. There are likely several examples of methodologies and tools that, in the wrong hands, may lead to harmful gain-of-function experiments or dual-use threats to public health or national security. A molecular biologist tinkering to improve genome editing tools would be wise to ponder the implications of the unethical use of the technology for human germline editing, as in the ‘CRISPR babies’ case3. Computational biologists using AI technology for drug discovery need to be aware that such models could be misused to design biochemical weapons4.

In their Comment, Lunshof and Rijssenbeek describe a model called ‘collaborative ethics’, which calls for life sciences researchers to work closely with professional ethicists or philosophers starting at the earliest stages of research planning. As they write, collaboration can “improve the efficiency and robustness of outcomes” for the research team and also “prepare a team for a formal ethics review and criticism after publication.” Such collaborations also benefit the field of philosophy, “as direct involvement with the sciences allows long-held assumptions and arguments to be put to the test.”

In the collaborative ethics model, a research team will discuss their ideas and concepts with a professional ethicist or philosopher, who may ask probing questions about the nature of the work. The ethicist will help the researchers understand whether there are any ethical considerations and whether the potential harms of sharing the resulting new knowledge with society outweigh the potential benefits of advancing science. Lunshof and Rijssenbeek highlight how the collaborative ethics model played a role in three different projects: the engineering of synthetic human entities with embryo-like features, the development of brain organoids and the programming of computer-designed ‘biobots’. They note that collaborative ethics may also have a role to play in many other fields, such as protein engineering, systems biology, aging research and computational biology.

As described in our journal policy, performing ethical research is the responsibility of all scientists. Professional ethicists are not the police; their role is not to punish or curb scientific progress. Rather, ethicists can serve as a ‘conscience’ and help scientists understand the ethical implications of their work. Without such collaborations, the after-the-fact consequences could be much worse — yes, papers can be retracted, but reputations are already damaged and harmful gain-of-function work may be out there for the world to see.

For scientists used to dealing with quantitative measures and statistical significance, the concept of research ethics can feel a bit vague or murky, at best an afterthought to their work. This is even more a reason for breaking down the walls between science and ethics, especially in this age of rapid technology development and especially with swift advances in AI. We strongly encourage institutions to go beyond minimal ethics training courses and do much more to support close collaborations between scientists and professional ethicists.