Zhou Xiaoqin, one of He Jiankui's colleagues, places an embryo in its storage tube in Shenzhen.Mark Schiefelbein / AP

Last year, Jennifer Doudna, one of the pioneers of the gene-editing technique known as CRISPR, said, “I’ve mentally prepared myself for the day when I open my inbox or answer my phone and realize that somebody’s going to be announcing the first CRISPR baby.”

That day was yesterday.

As first reported by Antonio Regalado at MIT Technology Review, the Chinese scientist He Jiankui claims to have made the first CRISPR-edited babies. “Two beautiful little Chinese girls, Lulu and Nana, came crying into the world as healthy as any other babies a few weeks ago,” He said in the first of five videos, posted yesterday to YouTube. “The girls are home now with their mom, Grace, and dad, Mark.” The claim has yet to be formally verified, but if true, it represents a landmark in the continuing ethical and scientific debate around gene editing.

Late last year, He reportedly enrolled seven couples in a clinical trial, and used their eggs and sperm to create embryos through in vitro fertilization. His team then used CRISPR to deactivate a single gene called CCR5 in the embryos, six of which they then implanted into mothers. CCR5 is a protein that the HIV virus uses to gain entry into human cells; by deactivating it, the team could theoretically reduce the risk of infection. Indeed, the fathers in all eight couples were HIV-positive.

Whether the experiment was successful or not, it’s intensely controversial. Scientists have already begun using CRISPR and other gene-editing technologies to alter human cells, in attempts to treat cancers, genetic disorders, and more. But in these cases, the affected cells stay within a person’s body. Editing an embryo is very different: It changes every cell in the body of the resulting person, including the sperm or eggs that would pass those changes to future generations. Such work is banned in many European countries, and prohibited in the United States. “I understand my work will be controversial, but I believe families need this technology and I’m willing to take the criticism for them,” He said.

The next few days will test his resolve. Ethicists and watchdogs have already called his work “monstrous,” “unconscionable,” and “a grave abuse of human rights.” Even the Southern University of Science and Technology, where He is a professor, denounced it. It issued a statement that called the work a “serious violation of academic ethics and standards.” As reported by Alexandra Harney and Kate Kelland at Reuters, the university said it was unaware of the project and plans to immediately launch an investigation by an independent committee. He, the researcher, reportedly began a three-year stint of unpaid leave in February.

“Was this a reasonable thing to do? I would say emphatically no,” says Paula Cannon of the University of Southern California. She and others have worked on gene editing, and particularly on trials that knock out CCR5 as a way to treat HIV. But those were attempts to treat people who were definitively sick and had run out of other options. That wasn’t the case with Nana and Lulu.

“The idea that being born HIV-susceptible, which is what the vast majority of humans are, is somehow a disease state that requires the extraordinary intervention of gene editing blows my mind,” says Cannon. “I feel like he’s appropriating this potentially valuable therapy as a shortcut to doing something in the sphere of gene editing. He’s either very naive or very cynical.”

“I want someone to make sure that it has happened,” says Hank Greely, an ethicist at Stanford University. If it hasn’t, thatwould be a pretty bald-faced fraud,” but such deceptions have happened in the past. “If it is true, I’m disappointed. It’s reckless on safety grounds, and imprudent and stupid on social grounds.” He notes that a landmark summit in 2015 (which included Chinese researchers) and a subsequent major report from the National Academies of Science, Engineering, and Medicine both argued that “public participation should precede any heritable germ-line editing.” That is: Society needs to work out how it feels about making gene-edited babies before any babies are edited. Absent that consensus, He’s work is “waving a red flag in front of a bull,” says Greely. “It provokes not just the regular bio-Luddites, but also reasonable people who just wanted to talk it out.”

This milestone isn’t surprising. In 2015, Junjiu Huang of Sun Yat-sen University used CRISPR to edit human embryos, but only nonviable ones that could never have resulted in a live birth. A second Chinese team followed suit the year after. And last year, Shoukhrat Mitalipov of Oregon Health and Science University became the first to use the technique on viable embryos—but never actually implanted any of these into a woman.

Given such work, it seemed almost inevitable that someone would take the steps that He did. “On one level, this isn’t a surprise at all,” says Ellen Clayton, a professor of law and health policy at Vanderbilt University. “On another, this is ...” at which point she was at a lost for words, and heaved a big sigh. “This calls into question the possibility of control, and I think it will exacerbate public fear,” she added.

Societally, the creation of CRISPR-edited babies is a binary moment—a Rubicon that has been crossed. But scientifically, the devil is in the details, and most of those are still unknown.

CRISPR is still inefficient. The Chinese teams who first used it to edit human embryos only did so successfully in a small proportion of cases, and even then, they found worrying levels of “off-target mutations,” where they had erroneously cut parts of the genome outside their targeted gene. He, in his video, claimed that his team had thoroughly sequenced Nana and Lulu’s genomes and found no changes in genes other than CCR5.

That claim is impossible to verify in the absence of a peer-reviewed paper, or even published data of any kind. “The paper is where we see whether the CCR5 gene was properly edited, what effect it had at the cellular level, and whether [there were] any off-target effects,” said Eric Topol of the Scripps Research Institute. “It’s not just ‘it worked’ as a binary declaration.”

Others found little comfort in the assurances from He’s YouTube video. “The mirror analogy here would be if you and I as physicians developed a cold fusion reactor ... without ever having done nuclear reactor design, flipped the switch [without] any safety inspection or even checking if it produced power, then posted [a] vid on Instagram,” said Sandip Patel, an immunologist at the University of California at San Diego, on Twitter.

In the video, He said that using CRISPR for human enhancement, such as enhancing IQ or selecting eye color, “should be banned.” Speaking about Nana and Lulu’s parents, he said that they “don’t want a designer baby, just a child who won’t suffer from a disease that medicine can now prevent.”

But his rationale is questionable. Huang, the first Chinese researcher to use CRISPR on human embryos, targeted the faulty gene behind an inherited disease called beta thalassemia. Mitalipov, likewise, tried to edit a gene called MYBPC3, whose faulty versions cause another inherited disease called hypertrophic cardiomyopathy (HCM). Such uses are still controversial, but they rank among the more acceptable applications for embryonic gene editing as ways of treating inherited disorders for which treatments are either difficult or nonexistent.

In contrast, He’s team disabled a normal gene in an attempt to reduce the risk of a disease that neither child had—and one that can be controlled. There are already ways of preventing fathers from passing HIV to their children. There are antiviral drugs that prevent infections. There’s safe-sex education. “This is not a plague for which we have no tools,” says Cannon.

As Marilynn Marchione of the AP reports, early tests suggest that He’s editing was incomplete, and at least one of the twins is a mosaic, where some cells have silenced copies of CCR5 and others do not. If that’s true, it’s unlikely that they would be significantly protected from HIV. And in any case, deactivating CCR5 doesn’t confer complete immunity, because some HIV strains can still enter cells via a different protein called CXCR4.

Nana and Lulu might have other vulnerabilities. Two decades ago, geneticists identified people who naturally have two broken copies of CCR5, and are resistant to HIV without any obvious health problems. For that reason, some billed the gene as a “safe harbor” that could serve as a minimally risky test site for gene editing. But there are risks: Later studies showed that CCR5 deficiencies make individuals more susceptible to other infections such as West Nile virus and Japanese encephalitis, and more likely to die when they catch influenza—a disease that kills hundreds of thousands of people every year.  

It is also unclear if the participants in He’s trial were fully aware of what they were signing up for. The team’s informed-consent document describes their work as an “AIDS vaccine development project,” and while it describes CRISPR gene editing, it does so in heavily technical language. It doesn’t mention any of the risks of disabling CCR5, and while it does note the possibility of off-target effects, it also says that the “project team is not responsible for the risk.”

He owns two genetics companies, and his collaborator, Michael Deem of Rice University, holds a small stake in, and sits on the advisory board of, both of them. The AP’s Marchione reports, “Both men are physics experts with no experience running human clinical trials.”

In a talk last year, He mentioned Jesse Gelsinger, an American teen who died in a botched gene-therapy trial in 1999. His death had a chilling effect that set the field of gene therapy back for more than a decade. To avoid a similar fate, He urged scientists to move cautiously before editing the genome of embryos.

And yet, He did it anyway.

“I’m worried that the potential backlash to this could slow down uses of this technology that currently seem acceptable,” says Greely. “I’d hate to see this reckless act stop an eventual social consensus.”

We want to hear what you think about this article. Submit a letter to the editor or write to letters@theatlantic.com.