Gene Edited Babies: The Science, Ethics, and Controversy

Three gene-edited babies are known to exist. All three were born in China in 2018 and 2019, the result of a secretive experiment by biophysicist He Jiankui, who used CRISPR-Cas9 to alter embryos before implantation. The scientific community condemned the work almost universally, and He was sentenced to three years in prison. Since then, no country has authorized a repeat, but the underlying technology has advanced rapidly, the ethical debates have only intensified, and the question of when or whether someone will try again looms over the field.

What He Jiankui Actually Did

He Jiankui attempted to disable the CCR5 gene in human embryos using CRISPR-Cas9, aiming to confer resistance to HIV. The gene encodes a receptor the virus uses to enter immune cells, and a naturally occurring variant called CCR5-delta32 is found in a small percentage of people of European descent who appear resistant to certain HIV strains. The idea, on its surface, sounded straightforward. In practice, the experiment was riddled with problems. An expert review of He’s presented data revealed what researchers described as a troubling lack of basic medical ethics alongside insufficient understanding of genetics and gene editing.

1PLoS Biology. Gene-edited babies: What went wrong and what could go wrong

The edits He achieved did not actually replicate the naturally protective CCR5-delta32 variant. Instead, the embryos ended up with novel mutations whose effects were unknown. At least one of the twin girls born from the experiment was mosaic, meaning some of her cells carried the edit and others did not, rendering the intended HIV protection incomplete at best. The experiment targeted a condition that already has effective prevention strategies, meaning the risk-benefit calculation was upside down from the start. And the families involved were HIV-discordant couples (the fathers were HIV-positive) who were reportedly told little about the experimental nature of the procedure.

Why CRISPR Keeps Breaking More Than It Fixes in Embryos

The core problem with using CRISPR-Cas9 in human embryos is not that it fails to cut DNA at the intended spot. It usually does. The problem is what happens after the cut. CRISPR-Cas9 creates a double-strand break, essentially snapping both rails of the DNA ladder, and then relies on the cell’s own repair machinery to put things back together. In early embryos, that repair process goes wrong at startling rates.

One study examining CRISPR-edited human embryos found that about 46% of the double-strand breaks failed to undergo appropriate repair. The outcomes included large-scale deletions, chromosomal rearrangements, and even total loss of the targeted chromosome.

2Human Reproduction. Assessment of genome editing outcomes in human preimplantation embryos subjected to CRISPR-Cas9

Separate research found unintended editing outcomes in roughly 16% of analyzed embryo cells, with damage extending well beyond the target site, spanning regions of four to twenty kilobases, and in some cases causing gains or losses of entire chromosome segments.

3PubMed Central. Frequent loss of heterozygosity in CRISPR-Cas9-edited early human embryos

The damage is not limited to the intended target. Studies in animal models show that about 6% of editing events produce large structural variants when on-target and off-target sites are considered together, including deletions large enough to remove entire gene exons that were never meant to be touched.

4Nature Communications. CRISPR-Cas9 induces large structural variants at on-target and off-target sites in vivo that segregate across generations

Perhaps most alarming, researchers have documented that CRISPR-induced breaks can trigger a catastrophic mutational process called chromothripsis, in which a chromosome essentially shatters and is reassembled incorrectly. This was observed in a majority of sequenced cell lineages in one study and represents a form of on-target toxicity that the researchers say cannot be completely avoided in many genome editing applications.

5bioRxiv. Chromothripsis as an on-target consequence of CRISPR-Cas9 genome editing

These are not hypothetical risks discovered in contrived lab conditions. They emerge directly from the mechanism CRISPR uses: creating a clean break and hoping the cell patches it correctly. Early human embryos appear to have relatively immature DNA repair pathways, which makes this gamble particularly poor in exactly the context where germline editing would occur.

Mosaicism and Why It Haunts Every Embryo Experiment

Even when the intended edit is made cleanly, there is no guarantee it reaches every cell. If the edit happens after the one-cell stage, or if different cells repair the break differently, the resulting embryo becomes a patchwork of edited and unedited cells. This is mosaicism, and it has shown up in virtually every published embryo-editing experiment.

In a study that used CRISPR to edit rhesus macaque embryos as a disease model, researchers achieved a high editing rate in the embryo overall, but when they later examined individual blood cells from the resulting infant, half carried the intended edit and half retained the unmodified gene.

6PubMed Central. CRISPR/Cas9 editing of the MYO7A gene in rhesus macaque embryos to generate a primate model of Usher syndrome type 1B

For research purposes, a mosaic animal can still be useful. For a child, mosaicism means some organs or tissues carry the disease-causing gene and others do not, creating unpredictable health outcomes. It also complicates any claim that the edit has been “corrected,” since genetic testing of a few cells might miss the mosaicism entirely. This is one of the reasons that, even after years of refining CRISPR delivery timing and concentrations, no laboratory has demonstrated reliably uniform editing in human embryos.

When Would Germline Editing Be Medically Justified?

A common question is why anyone would edit an embryo when preimplantation genetic testing already lets fertility clinics screen embryos for inherited diseases and select unaffected ones. The answer is that embryo screening works well for most couples carrying genetic disease risk, but it has blind spots. If a parent is homozygous for a dominant disease like Huntington’s, every embryo they produce will inherit the mutation. The same is true when both parents are homozygous for a recessive condition like cystic fibrosis, or when a parent carries certain chromosomal rearrangements. In those rare cases, screening cannot find an unaffected embryo because none exists.

7PubMed Central. Germline genome editing versus preimplantation genetic diagnosis: Is there a case in favour of germline interventions?

Germline editing experiments in human embryos have primarily focused on correcting mutations linked to these kinds of inherited diseases.

8PubMed. The case for germline gene correction: state of the science

The clinical niche is genuine but narrow. The number of couples for whom embryo screening is truly impossible is small, and the option of using donor gametes can solve the problem for many of them, though some couples understandably prefer to have a genetically related child free of the disease. Whether that preference justifies the current risks of embryo editing is where the medical argument runs into the safety evidence described above.

Newer Editing Tools That Skip the Double-Strand Break

Much of the safety concern around CRISPR-Cas9 in embryos traces back to that double-strand break. Newer approaches try to avoid it entirely. Base editing chemically converts one DNA letter to another at a precise location without cutting both strands. Prime editing uses a modified Cas9 fused with a reverse transcriptase to rewrite short stretches of DNA, also without a full break.

In mouse embryos, a prime editing system called PEmbryo achieved an average of 58% precise editing across multiple target sites, with on-target errors at just 0.5%. Whole-genome sequencing revealed some off-target insertions and deletions in repetitive regions of the genome, but no obvious harmful effects in the resulting mice.

9Nature Biotechnology. Efficient prime editing in two-cell mouse embryos using PEmbryo

Base editing has also been tested directly in human embryos. Recent work with adenine base editors showed that the approach did not produce the large deletions or chromosomal damage characteristic of standard CRISPR-Cas9. However, mosaicism remained common, and off-target editing varied depending on the guide RNA used. The researchers concluded that base editing appears substantially less genotoxic than CRISPR-Cas9 in human embryos but is not ready for clinical use due to unresolved mosaicism and off-target activity.

These tools represent genuine progress. They narrow the gap between what researchers can do safely and what would be needed for clinical application. But “less genotoxic” is not the same as “safe enough for a baby,” and neither base editing nor prime editing has solved mosaicism, which remains the single most stubborn technical obstacle.

The Consent Problem

Beyond the technical hurdles, germline editing raises a philosophical question that no amount of engineering can resolve: the edited person never consented to being edited. Unlike somatic gene therapy, where a patient chooses treatment for themselves, germline changes are imposed on someone before they exist and then passed to their children and grandchildren.

This argument carries real weight. The U.S. National Institutes of Health explicitly cited the inability to obtain consent from future generations as a reason for not funding germline editing research. But critics of the consent argument point out that we routinely make irreversible decisions that shape the lives of people who cannot consent to them. Parents choose where to live, what to eat, whether to vaccinate, all of which profoundly affect a future child. The central question, as one philosophical analysis put it, is not whether the individuals exposed to the risks would consent, but whether they will also enjoy benefits that outweigh those risks.

10PubMed Central. The Ethics of Germline Gene Editing

The analogy has limits. Choosing a neighborhood is reversible in a way that altering someone’s DNA is not. And while every reproductive decision shapes a child’s genome through the lottery of recombination, actively rewriting specific genes feels qualitatively different to many ethicists. The debate is far from settled, but it has become more nuanced than the early framing of “playing God” suggested.

Disability Rights and the Message of Elimination

Disability rights advocates have raised a concern that often gets lost in the excitement over disease prevention: if society develops the capacity to edit out genetic conditions, what does that say about people currently living with those conditions? Research exploring the perspectives of people with genetic conditions and their families has found significant worry that genome editing technologies carry an implicit message that disabled lives are of lesser value.

11PubMed Central. Human genome editing and the identity politics of genetic disability

This is not an abstract concern. Deaf communities, for instance, have a rich cultural identity and a long history of resisting medical frameworks that treat deafness as a deficiency to be corrected. Some genetic conditions that cause chronic illness also come with traits their carriers value. The line between “treating disease” and “eliminating human variation” is not as obvious as it seems from outside these communities, and the conversation about germline editing has often proceeded without meaningful input from the people most directly affected.

Who Could Afford It

If germline editing ever becomes safe enough for clinical use, it will almost certainly be expensive. Existing gene therapies already cost between $450,000 and $2 million per treatment, with some reaching as high as $3.5 million for a single dose. These prices place gene therapies primarily within the reach of society’s most advantaged while excluding much of the population, particularly individuals from historically disadvantaged groups.

12PubMed Central. CRISPR in Public Health: The Health Equity Implications and Role of Community in Gene-Editing Research and Applications

The equity concern runs deeper than sticker price. If germline editing could eventually prevent serious genetic diseases, and only wealthy families could access it, the technology would not merely fail to close health disparities but would actively widen them. Children born to families who could afford editing would carry biological advantages their peers could not access. This dynamic tracks what public health researchers call fundamental cause theory: every major medical advance tends to benefit the socially advantaged first, because they have the resources to learn about, access, and pay for new interventions.

Some researchers have pointed out that this is not unique to gene editing. IVF itself was once a luxury available only to the wealthy and is still unevenly accessible worldwide. But the permanence of germline changes raises the stakes. A wealthy person who gets IVF gains a child; a wealthy person who gets germline editing gains a child whose genetic advantages propagate through every subsequent generation.

The Regulatory Landscape

As of now, every nation that has weighed in agrees that heritable genome editing should not proceed clinically.

13PubMed. Global Governance of Human Genome Editing: What Are the Rules?

The mechanisms behind that consensus vary enormously. In the United States, a congressional rider has prohibited the FDA from even reviewing applications for clinical trials involving heritable genetic modifications, which effectively bans the procedure without any formal regulatory decision on its merits. The United Kingdom takes a more structured approach, with the Human Fertilisation and Embryology Authority overseeing reproductive technologies and Parliament having to change the law before clinical germline editing could be approved. China, where the He Jiankui experiment took place, has updated multiple laws and regulations in the aftermath, though analyses of the Chinese framework still identify shortcomings in enforcement and oversight.

14PubMed Central. Regulatory framework of human germline and heritable genome editing in China: a comparison with the United States and the United Kingdom

The fragility of this consensus is the real concern. If one country were to demonstrate a safe and effective germline editing procedure, the global agreement to wait could fracture quickly. Countries with lighter regulatory burdens or more permissive cultures around reproductive technology could become destinations for germline editing, much as some already serve as hubs for surrogacy or experimental stem cell therapies.

Rogue Clinics and Fertility Tourism

The He Jiankui case demonstrated that a single determined researcher with access to IVF infrastructure and CRISPR reagents could carry out germline editing largely outside any oversight system. Experts have warned that this pattern could repeat, and potentially worsen, as the technology becomes cheaper and more accessible. Different forms of regulatory oversight, combined with differences in clinical cultures and patient needs, will probably result in additional premature applications, according to a policy analysis that also flagged the likely emergence of commercial clinics offering germline editing for nontherapeutic purposes.

15PubMed Central. Heritable Genome Editing in a Global Context: National and International Policy Challenges

Illegal and rogue clinics operating outside research protocols would expose patients, embryos, and any resulting children to significant, unjustified risks. The fertility industry already has experience with medical tourism: patients travel across borders for procedures banned or unavailable in their home countries. Adding germline editing to that ecosystem is not a matter of if but of what regulatory structures can do to delay and contain it.

What the Public Actually Thinks

Public opinion on gene-edited babies is more divided than expert opinion. After the He Jiankui announcement, researchers tracked reactions across multiple social media platforms and found a stark gap between professional scientists and the general public. While almost all experts opposed the experiment, many online posts supported it. Opposition was highest on Twitter, where roughly 86% of posts were critical, and lowest on YouTube, where opinion was essentially split, with about 53% opposing. The primary reason for opposition was ethical concern, while the primary reason for support was the hope that such technology could prevent future diseases.

16PubMed Central. The Public Perception of the #GeneEditedBabies Event Across Multiple Social Media Platforms: Observational Study

The gap makes sense when you consider the information asymmetry. Scientists who read the data saw an experiment with poor execution, questionable justification, and alarming safety signals. Members of the public who saw headlines saw a technology that might one day eliminate genetic diseases in children. Both reactions are rational given the information each group had access to, which underlines why public engagement on this topic needs to go beyond yes-or-no opinion polls and actually communicate the specific risks that make the scientific community so cautious.

Epigenome Editing as a Possible Middle Path

An approach that has received less public attention involves editing the epigenome rather than the genome itself. Instead of changing the DNA sequence, epigenome editing alters chemical tags, primarily DNA methylation patterns, that control whether genes are turned on or off. In theory, this could silence a disease-causing gene without permanently rewriting the genetic code, and the modifications might not be heritable in the same way as sequence changes.

17PubMed Central. Editing DNA Methylation in Mammalian Embryos

The research is still early. Mammalian embryos undergo dramatic reprogramming of their methylation patterns during normal development, which means an epigenetic edit introduced at one stage might be erased naturally at the next. Understanding when and how to intervene in that process without disrupting normal development is an active area of investigation. But if the approach matures, it could offer a way to modulate gene expression in embryos without the irreversible DNA sequence changes that make germline editing so fraught. Whether regulators and ethicists would treat epigenome editing differently from genome editing remains an open question, since the intent and functional outcome could be identical even if the molecular mechanism differs.

Mitochondrial Replacement as a Regulatory Rehearsal

One underappreciated part of this story is that a form of heritable genetic intervention already exists. Mitochondrial replacement therapy, approved in the United Kingdom in 2015, substitutes a mother’s defective mitochondrial DNA with healthy mitochondria from a donor. The resulting child inherits genetic material from three people, and that change passes to future generations through the maternal line. The procedure has been described as a crucial test case and learning guide for the scientific, ethical, and regulatory challenges of future reproductive breakthroughs.

18PubMed. Going Germline: Mitochondrial Replacement as a Guide to Genome Editing

The regulatory process for mitochondrial replacement took over a decade of review in the UK alone. It involved extensive public consultation, parliamentary debate, and ongoing monitoring of children born through the procedure. If germline genome editing ever approaches clinical readiness, that deliberate and transparent model is the closest template the world has for how to manage the transition. Whether any country will show similar patience when CRISPR-based approaches mature is another question entirely.