“The unthinkable has become conceivable,” said David Baltimore from the California Institute of Technology, at a historic summit on human gene editing currently taking place in Washington, D.C. “We are close to altering human heredity and we need to decide how we as a society are going to use this capability.”

The summit—a three-day event organized by august scientific institutions from three countries—offers a chance for scientists, ethicists, lawyers, and interested members of the public to “consider the scientific and societal implications of genome editing” at a time when it has never been easier or more powerful. It’s a spiritual successor to a similar conference at Asilomar, California, in 1975, when delegates debated the ethics of nascent genetic-engineering technology.

Baltimore was involved in both meetings, and he says that things are very different now. The difference lies in a suite of new tools for changing a person’s DNA, especially the much-hyped CRISPR-Cas9 system, which allows scientists to easily delete, tweak, or insert genes. And compared to other older techniques, it’s cheaper, more precise, and more versatile. It’s like a computer next to an abacus or, if you have a bleaker outlook, a machine-gun next to a spear.  

With this power at hand, old questions about playing God, making designer babies, and ushering in dystopian Brave New Worlds of genetic haves and have-nots, take on fresh urgency. These same leitmotifs are trotted out with every new wave of genetic technology—IVF, cloning, stem-cell therapies, mitochondrial-replacement therapy—but some say they are more pertinent than ever. “In the past, it’s been simple for scientists to dismiss these possibilities,” said Robin Lovell-Badge from The Francis Crick Institute. “But we’re rapidly getting to the point where we can no longer deny them.”

This turning point came in April, when a team led by Junjiu Huang from Sun Yat-sen University, announced that they had edited human embryos with CRISPR, tweaking the faulty gene behind an inherited disease called beta-thalassaemia. (The team used inviable embryos that could never have developed into an actual person.) The research caught the world off guard. In the U.S., it could not have been done, at least not with federal funding. In the U.K., it’s permissible with a license, as long as embryos are younger than 14 days.

In response, some scientists have called for a moratorium on any kind of human germline editing until the world can assess the safety of CRISPR and other gene-editing tools, discuss the social consequences of such technologies, and draw up clear ethical guidelines and regulations.

Hence the summit. But the delegates spent much of the first day wrestling with a more subtle question: Even if we could edit human genes safely and precisely, why would we do so?

If you’re looking to edit genes in actual people, the most obvious application is to treat diseases. A month ago, doctors cured a girl with untreatable leukaemia by removing immune cells, editing them so they’d go after cancer cells while also resisting a chemotherapy drug, and then injecting them back into her. Other teams are trying to remove cells from people with HIV, deleting a gene that the virus needs to stage its invasions, and injecting them back in. These are examples of somatic cell therapy—they affect cells that stay within a person’s body and die with them.

It’s far more controversial to edit the human germline—that is, genes of sperm, eggs, or early embryos. These modifications wouldn’t just affect one individual but also their descendants. They would cascade down generations, potentially altering the course of human heredity.

Germline modifications would make the biggest difference in cases of severe inherited diseases, like cystic fibrosis, Huntington’s disease, or Tay-Sachs disease, which all cause debilitating symptoms, carry a poor prognosis, and are caused by mutations in single genes. In an era with safe, efficient gene-editing, these conditions could be entirely preventable. “I’m the mother of a child who died because of a fetal birth defect,” said Sarah Gray from the American Association of Tissue Banks to one group of panelists. “He was six days old and he suffered every day. He had seizures every day. If you have the skills and the knowledge to fix these diseases, then frickin’ do it.”

There are many other common diseases, including Alzheimer’s, diabetes, and several cancers, which are under the influence of hundreds or thousands of genes. In these cases, editing genes would be a way of lowering risk, rather than outrightly preventing disease.

And perhaps most controversially of all, people could conceivably use CRISPR and other gene-editing tech to make non-medical enhancements, like tweaking genes that increase muscle mass, or height, or intelligence. That’s where the familiar refrains about designer babies comes in.

That’s the big picture. Now for some practicalities.

Consider the severe genetic disorders, like Tay-Sachs or Huntington’s. In dominant genetic diseases (where one copy of a faulty gene is enough to cause the condition), 50 percent of embryos will be normal; in recessive genetic diseases (where two faulty copies are needed), 75 percent will be normal. So most won’t require any editing at all. Parents could opt for in-vitro fertilization (IVF), after which doctors could screen the resulting embryos for those that don’t carry any copies of the risky genes. This technique, known as pre-implantation genetic diagnosis (PGD), only fails when both parents carry the necessary variants, and all their embryos would be similarly affected. Those pairings, however, are very rare.

“If we really care about avoiding genetic diseases, germline editing isn’t the first, second, third or fourth thing that we should be thinking of,” said Eric Lander from the Broad Institute. Instead, it would do more good to make genetic diagnostic tests more widely available, so parents would know that they were carriers of risky genes, and could sign up for PGD.

But PGD isn’t always efficient. George Daley from Harvard Medical School has treated several children affected by NEMO deficiency syndrome, where a faulty gene leaves them with a terribly weak immune system. Families often try to have a second healthy child, whose bone marrow can be used to save their older sibling. PGD seems like an obvious solution, but couples tend to be older at that point and success rates are low. In the last year, Daley’s colleagues have tried this approach for eight families, and though each tried an average of five IVF rounds, only one conceived an actual baby. In situations like this, says Daley, it might be better to use CRISPR to make decisive edits to a single embryo, rather than to play the lottery and hope to find a viable one.

George Church, a CRISPR pioneer from Harvard University sees possible paths from these legitimate uses to more dubious ones. “I think enhancement will creep in the door in terms of treating serious diseases,” he says. Someone who is losing their faculties to Alzheimer’s disease might turn to gene-editing to stem their cognitive decline. “Then, someone younger with a super high risk of the disease. Then, a business executive who wants to get ahead of the game. Then that same executive who wants to fix his sperm cells.”

But gene-editing might be totally impractical for fixing common diseases, because they are typically influenced by legions of genes. If you took people with the highest risk of, say, schizophrenia, you’d probably need to CRISPR thousands of genes to bring their odds back down to average levels. That’s a terrible idea, for reasons we’ll get to.

The same is true for attributes like intelligence, height, sporting ability, or personality traits, which involve small contributions from thousands of genes, and a massive dollop of environmental influence on the side. “The dishes do not come à la carte,” writes Nathaniel Comfort, a historian of medicine at Johns Hopkins University in Baltimore. “If you believe that made-to-order babies are possible, you oversimplify how genes work.”

So, no matter how precisely we can edit genes, there are some things we won’t be able to edit our way out of—and certainly not safely. Genes rarely do one thing. Those thousands of edits will ripple through the body in unexpected ways. For example, deleting the CCR5 gene would make people resistant to HIV, but also make them 13 times more likely to die of West Nile virus. Tweaking their FUT2 gene might make them less likely to develop Type 1 diabetes but also make them vulnerable to norovirus.

Even genes behind severe diseases can have unexpected benefits. The variant that causes sickle-cell anaemia has stuck around in the human population because it also protects carriers against malaria. Removing that variant from the human gene pool might not be a great idea.

“I can only think of a handful of things that are plausible variants for editing,” says Lander. For example, people with the E4 version of the ApoE gene have much higher risks of Alzheimer’s disease. You could edit that, says Lander, “but I can’t swear there’d be no problem because ApoE4 has been kept around in 3 percent of every human population.” That gives him pause. “If [editing that gene] is such a good idea, why didn’t evolution think about doing it?” he asks.

The risk is that some people won’t care about any of this. “Any uses of new technology are partly driven by need, but we shouldn’t forget that they’re also driven by demand,” Daley said. With CRISPR being easy to both use and misuse, there’s a chance that unscrupulous clinics will offer gene-editing enhancements, even if scientists deem them unsafe, ineffective, or imaginary. After all, hokey vitamin supplements make rip-roaring trade, as do unproven and unregulated stem-cell therapies. If these cases tell us anything, the answer to “Why would we use this technology?” might well be “Because we can.”