- The Core Ethical Distinction: Somatic vs Germline
- He Jiankui: When the Line Was Crossed
- International Governance: Who Makes the Rules?
- The Treatment vs Enhancement Debate
- Consent: The Problem of Future Persons
- Equity: Who Gets Access to CRISPR Cures?
- Gene Drives: Ecological Ethics at Scale
- The 20-Year Horizon: What Is Coming
- A Closing Reflection on Responsibility
- The Final Story: The Author of the Code
Section 1 — The Core Ethical Distinction: Somatic vs Germline Editing
The most important ethical line in all of gene editing — the one that almost all scientists, ethicists, and regulatory bodies agree on as a bright line — is the distinction between somatic editing and germline editing. Everything else in the ethics of CRISPR follows from this distinction.
Somatic editing modifies the cells of a living person in a way that affects only that person. The edits are not heritable — they die with the patient. Casgevy edits haematopoietic stem cells; those cells and their progeny carry the edit, but the patient’s children will not. This is identical in principle to any other medical intervention. The patient consents, the risks and benefits affect only them, and the ethics are governed by the same framework as any other experimental therapy: informed consent, benefit-risk assessment, regulatory oversight, and post-market surveillance.
Germline editing modifies embryos, eggs, or sperm. Changes made at this stage are present in every cell of the resulting person — including their own reproductive cells, which means the change is heritable and will be passed to all future descendants. This is a categorically different ethical situation. The person who will carry the edit for life has not consented — they cannot consent, because they do not yet exist. Their children cannot consent. Their children’s children cannot consent. A germline edit is a permanent change to a family lineage and, if it spreads widely, to the human gene pool itself.
- Affects only the treated individual
- Not heritable — cannot be passed to children
- Patient can consent as an adult
- Regulated by standard medical/drug frameworks
- Currently approved (Casgevy) or in clinical trials
- Affects every cell of the resulting person
- Heritable — passed to all future descendants
- No consent possible from future person
- No established regulatory framework
- Currently banned or under strict moratoriums in most countries
Section 2 — He Jiankui: When the Line Was Crossed
In November 2018, a 34-year-old Chinese biophysicist named He Jiankui stood before an audience at the Second International Summit on Human Genome Editing in Hong Kong and announced that he had created the world’s first gene-edited human babies. Twin girls, named Lulu and Nana (pseudonyms), had been born weeks earlier with CRISPR edits in the CCR5 gene — a gene encoding a co-receptor that HIV uses to enter CD4+ T cells. He claimed the edit was intended to protect the girls from HIV infection.
The room went silent. Then it erupted. The announcement was not received as a scientific triumph. It was received as a catastrophic ethical violation, and the subsequent global response was swift and unequivocal: He Jiankui had conducted dangerous, premature, medically unjustified human experiments on embryos without proper informed consent, without ethical oversight, and without scientific justification. He was subsequently sentenced to three years in prison by a Chinese court for “illegal medical practice.”
Why the CCR5 Experiment Was Indefensible
The scientific community’s condemnation was not motivated by squeamishness about gene editing. It was rooted in specific, well-founded objections:
The girls’ father was HIV-positive, but mother-to-child and standard sperm-washing techniques would have prevented any meaningful transmission risk. There was no medical need that justified exposing the girls to the unknown risks of germline editing.
Loss of CCR5 function is associated with increased susceptibility to West Nile virus and influenza. The very mutation He was introducing — the CCR5-Δ32 variant — is known to increase mortality from certain flavivirus infections. He traded a hypothetical HIV protection for a real increased risk of other infections.
Subsequent analysis of the published data revealed that one of the twins (Nana) was mosaic — not all cells carried the edit. The intended protection was therefore incomplete. The girls received the risks of germline editing without the purported benefit.
Parents in the trial were recruited from an HIV advocacy group and were told they were enrolling in an ‘AIDS vaccine’ trial. The true nature of the experiment — irreversible germline editing of their children — was not adequately explained. The consent forms were misleading.
The experiment was conducted in secret, outside any established ethics review framework, without publication in a peer-reviewed journal before the announcement. The results were shared first at a conference and on YouTube.
“This work represents a serious breach of scientific norms. We call on the scientific community to vigorously oppose any attempt to use these technologies to alter the human germline until it has been deemed, after broad societal consensus, to be safe and ethically justified.”
The He Jiankui case did not end the conversation about germline editing — it clarified it. There is broad scientific agreement that germline editing is not currently safe or ethical, but not that it could never be. The question is not “never” but “not yet” — and the conditions under which “not yet” becomes “with appropriate safeguards” are the subject of ongoing international discussion.
Section 3 — International Governance: Who Makes the Rules?
Gene editing is a global technology, and the most important governance challenge it presents is not regulatory within any single country, but international: how do you prevent a race to the least regulated jurisdiction? If germline editing is banned in the US, EU, and China but permitted in country X, nothing prevents a well-funded actor from travelling to country X and proceeding. This is not hypothetical — it is exactly what He Jiankui did.
The International Summit Process
The major scientific academies of the US, UK, and China co-organised the first International Summit on Human Gene Editing in Washington DC in 2015, three years after the Doudna-Charpentier paper. The summit’s statement declared that it would be “irresponsible to proceed with clinical applications of germline editing unless and until the safety issues have been resolved, and there is broad societal consensus about the appropriateness of the proposed application.” A second summit in Hong Kong in 2018 was the one He Jiankui gate-crashed. A third summit in London in 2023 reviewed progress and reiterated that germline editing for reproduction remains premature.
The WHO Expert Advisory Committee
In 2021, the WHO published two landmark documents: a governance framework and a recommendations paper for human genome editing. Key recommendations included: the establishment of a global registry of all human genome editing research, enhanced mechanisms for information sharing and oversight, attention to equity of access, and a clear statement that it would be “irresponsible at this time to proceed with clinical applications of human germline genome editing.”
The WHO framework is advisory, not binding. It has no enforcement mechanism. Countries can choose to follow it or not. Its value is in establishing international scientific consensus, creating reputational costs for non-compliance, and providing a model that national regulatory bodies can adopt into binding domestic law. As of 2025, over 70 countries have some form of restriction on germline editing, ranging from complete legislative bans to voluntary moratoriums in the research community.
Section 4 — The Treatment vs Enhancement Debate
Even when germline editing is eventually considered safe enough to discuss clinically, it faces a deeper philosophical divide: the line between treatment and enhancement. This is not a new debate — it predates CRISPR — but CRISPR makes it urgent in a way that was previously hypothetical.
Treatment is using a medical intervention to bring a person to a normal or average state of health. Correcting the sickle cell mutation restores haemoglobin to normal function. Correcting a severe combined immunodeficiency gives a child an immune system they would normally have. These are interventions with clear medical need, clear targets, and clear benefit-risk calculus.
Enhancement is using an intervention to give a person capabilities beyond the typical range. Editing genes associated with greater intelligence, taller stature, lower disease risk, or greater athletic performance in healthy individuals falls in this category. The distinction sounds clear, but the boundary is genuinely fuzzy at the edges.
The Hard Cases at the Boundary
The APOE4 allele dramatically increases Alzheimer’s risk. Editing APOE4 to APOE3 in an embryo would reduce late-life dementia risk. Is this treatment (preventing a disease) or enhancement (improving a trait beyond baseline)? If a person would have had Alzheimer’s without the edit, it looks like treatment. But the person has not yet developed Alzheimer’s; they simply have a genetic risk factor that most people do not have. This case genuinely falls at the boundary.
The Deaf community raises a profound challenge to the treatment framing: many Deaf people do not consider deafness a disease to be cured, but a difference that constitutes part of their cultural identity. To them, prenatal editing to prevent deafness is not treatment — it is the erasure of a form of human diversity. This is not a fringe view; it is a widely held position within a community that has developed rich language, culture, and institutions around a trait that medicine classifies as a disorder.
He Jiankui argued his CCR5 edits were treatment — protecting children from a disease their father had. Critics argued it was neither: the children had no HIV infection to treat, and the risk to them was not meaningfully higher than the general population given their mother’s negative status and available preventive measures. The edit gave a hypothetical benefit (partial HIV resistance) at the cost of a real risk (increased flavivirus susceptibility). This illustrates how the treatment vs enhancement framing can be used rhetorically to justify almost any intervention.
Section 5 — Consent: The Problem of Future Persons
The consent problem in germline editing is genuinely novel. Medical ethics requires informed consent as its foundation: a patient must understand what is being proposed, understand the risks and benefits, and voluntarily agree. In germline editing, the entity most affected by the intervention — the future person who will live with the edit for their entire life — cannot be consulted.
This is sometimes compared to other decisions made for children before they can consent: vaccination, circumcision, dietary choices, education. But germline editing is different in degree in a way that may constitute a difference in kind. A vaccination affects the immune system during childhood and can be discussed with the now-adult patient if concerns arise. A germline edit affects every cell in the body from conception, is permanent and heritable, and may have effects that only manifest decades later. There is no way to “undo” a germline edit once the person has been born.
“The issue of consent is not merely procedural. It reflects a deep question about whether we have the right to make permanent biological choices on behalf of persons who do not yet exist, who cannot evaluate the decision, and who cannot reverse it.”
Some bioethicists argue that the “open future” principle — the idea that children should be raised in ways that maximise their future autonomy — places strong constraints on germline editing even for therapeutic purposes. A child edited to lack CCR5 receptors has had a permanent decision made about their biology, with implications for their health and identity, that they had no say in. The question of whether parents can legitimately make this kind of decision on behalf of future children is genuinely contested.
Section 6 — Equity: Who Gets Access to CRISPR Cures?
The $2.2 million price of Casgevy is not just an economic fact. It is an ethical statement about who gene editing is for. Sickle cell disease affects approximately 300,000 children born annually with the severe SS genotype, 80% of them in sub-Saharan Africa. Casgevy is not available in Africa. It is not approved there. It is not manufactured there. The health systems of most African countries could not afford it even if it were.
This creates a profound equity problem: the disease burden falls most heavily on populations with the least access to the cure. This is not unique to CRISPR — it is a structural feature of how pharmaceutical innovation is incentivised and distributed globally. But it is particularly acute for CRISPR because the technology was explicitly developed with sickle cell disease as a showcase application, and sickle cell disease is explicitly a disease of African populations.
The Structural Roots of Inequity
The high price of Casgevy reflects genuine manufacturing costs (each batch is custom-made for one patient), the enormous investment in clinical development (hundreds of millions of dollars over a decade), and the expectation of profit on that investment. In a market-based pharmaceutical system, this pricing is rational from a business perspective. But the market optimises for willingness to pay, not need — and the populations with the greatest need for sickle cell treatment have the least ability to pay.
Several potential solutions are in development. In vivo approaches (delivering CRISPR directly into patients rather than through complex ex vivo cell manufacturing) could dramatically reduce cost per patient. The Gates Foundation’s investment in low-cost sickle cell therapy aims for a target of under $100,000 per patient — still not affordable for most of sub-Saharan Africa, but orders of magnitude lower than Casgevy. Technology transfer to developing country manufacturers (similar to the COVID-19 vaccine technology transfer discussions) is another route being explored. None of these have yet been realised at scale.
Section 7 — Gene Drives: Ethics at Ecological Scale
The ethics of gene drives is distinct from the ethics of medical or agricultural CRISPR in one crucial respect: the scale of potential consequence. A gene drive, once released into a wild population, cannot be recalled. It spreads autonomously through an entire species, across borders, across generations. The ecological consequences of eliminating or radically altering a wild species cannot be fully predicted. And the communities most affected by both the disease (malaria) and the intervention (mosquito population suppression) are the same communities — in West Africa — that have historically had the least power in global scientific and regulatory decision-making.
The Four Ethical Principles That Gene Drives Must Satisfy
Regulatory approval from a national government is necessary but not sufficient. The communities living alongside the mosquito population — who will experience both the potential benefit and any unexpected ecological consequences — must have genuine input into the decision. Target Malaria’s decade-long community engagement in Mali and Burkina Faso is a model, though critics argue even this does not constitute true consent for an irreversible intervention.
600,000 deaths per year from malaria is not a small number. The proportionality calculation must weigh this human cost against the ecological risk of mosquito population suppression. A drive that reduces malaria mortality by 90% while causing modest ecological disruption in one insect species may be ethically justifiable. A drive that reduces mortality by 10% while causing cascading ecosystem collapse would not be. The evidence for each must be assessed rigorously before any release.
Before any suppression drive is released, a reversal drive that could restore the wild-type population should be developed and stored in laboratory containment. This does not make the decision reversible, but it provides an option that would not otherwise exist. The technical feasibility of reversal drives for Anopheles has been demonstrated in principle; their practical deployment in the field has not.
The first release should not be in an entire continent. It should be geographically contained, intensively monitored, with clear criteria for proceeding to wider release or halting. Island populations, peninsulas, or isolated valleys could provide containment while allowing real-world data to be collected. This phased approach allows learning before irreversible commitment.
Section 8 — The 20-Year Horizon: What Is Coming
Predicting technological development is notoriously unreliable. In 2012, essentially no one predicted the pace at which CRISPR would move from bacterial immune system to approved human therapy. With that caveat firmly in mind, here is a credible assessment of what the next two decades may bring — based on trajectories already visible in current research.
Multiple CRISPR cell therapies approved for blood diseases, some cancers, and possibly HIV. Manufacturing costs begin to fall as more facilities are built and processes are optimised. First base editing and prime editing therapies enter Phase 2/3 trials. Pig-to-human organ transplants move from compassionate use to Phase 2 clinical trials. CRISPR-edited crops commercially grown in US, Japan, Brazil, and increasingly Asia.
LNP delivery to tissues beyond the liver (muscle, lung, possibly CNS) becomes feasible. First in vivo Duchenne MD therapy approved. CRISPR approaches for Huntington’s and ALS enter Phase 2. Gene drive field trials begin in contained geographical settings for malaria. CRISPR-edited livestock enter commercial supply chains in multiple countries. Germline editing debate intensifies as safety data accumulates in somatic applications.
CRISPR therapies available for dozens of single-gene disorders. Manufacturing cost reductions make ex vivo therapies accessible in middle-income countries. In vivo editing moves from rare disease to common disease prevention (PCSK9 for cardiovascular disease). Whole-genome sequencing at birth followed by preventive CRISPR editing discussed seriously in some healthcare systems. Epigenome editing matures into a therapeutic modality. Gene drive field releases for malaria evaluated on evidence from first trials.
Twenty years of somatic editing safety data will have accumulated. The technical barriers to safe germline editing — mosaicism, off-target effects, delivery — may have been substantially addressed. At this point, the “not yet” framing begins to require a different answer. Whether that answer is “yes, under strict conditions for severe diseases” or “still no, the consent and equity problems are not solved” will depend on political and ethical choices that have not yet been made.
Section 9 — A Closing Reflection on Responsibility
There is a pattern in the history of powerful technologies: the people who develop them rarely fully anticipate the uses to which they will be put. The physicists who split the atom were not anticipating Hiroshima. The computer scientists who designed the internet were not anticipating social media’s effects on democracy. Jennifer Doudna has spoken publicly about waking in a cold sweat having dreamed that she explained CRISPR to a figure who turned around to reveal Adolf Hitler’s face. The dream was not prophetic, but it reflects the weight that the technology’s creators feel about its potential for misuse.
That weight is appropriate. CRISPR is not inherently dangerous — it is inherently powerful, which means it is dangerous if misused and transformative if used well. The history of medicine, of agriculture, of technology is a history of humans learning to use new powers responsibly — sometimes slowly, sometimes after terrible mistakes, but overall in a direction that has improved human health and welfare. There is no reason to think CRISPR will be different, provided we take the governance questions as seriously as we have taken the technical ones.
What the CRISPR story shows, above all, is that the most important decisions about powerful technologies are not made by the scientists who develop them. They are made by the societies that adopt, regulate, fund, and define the boundaries of their use. Yoshizumi Ishino could not have imagined in 1987 that a footnote in his bacterial genetics paper would lead, 36 years later, to an approved therapy that cures sickle cell disease. Nor can anyone today fully anticipate what the decisions made in the next decade about germline editing, gene drives, and enhancement will lead to in 2060.
“We are all now living in the consequences of a revolution whose full implications we do not yet understand. The only honest thing to do is to say so, to govern carefully, and to keep asking the hardest questions.”
The Author Who Learned to Rewrite the Book
In the beginning, the Book was given. No single author wrote it — it was written by four billion years of copying errors and selection, trial and catastrophic failure, extinction and adaptation. Every living thing on Earth is a version of the same book, endlessly revised across the generations. The Book is the genome. Its alphabet has four letters. Its language encodes everything that lives.
For most of the history of the Book’s existence, no creature could read it, let alone rewrite it. The revisions were made by forces that had no intention: ultraviolet light, copying errors, the random rearrangements of evolution. Then, in the last geological instant — a blink in the four-billion-year history of life — one species learned to read the Book. And then, in an even smaller instant, it learned to write in it.
The first edits were clumsy: restriction enzymes that cut DNA at fixed sequences, ligate foreign genes in, and hope the result worked. The second generation was better: zinc finger nucleases, TALENs — engineered proteins that could find specific sequences and cut. These were like learning to erase and redraw specific words in the Book, but the pencil was expensive, the work was slow, and only a handful of laboratories in the world could afford the tools.
Then CRISPR arrived. Suddenly, every laboratory in the world had a pencil. Not an expensive, custom-forged pencil that took years to make — a twenty-nucleotide RNA that could be ordered for a few dollars and changed in an afternoon. The Book was now writable by anyone with basic molecular biology skills. This was the democratisation of authorship, and it changed everything.
The new authors have done extraordinary things. They have corrected a single wrong letter in the haemoglobin chapter that had been causing suffering for millennia, and people have walked out of hospitals cured. They have strengthened the disease-resistance chapter of wheat, and crops that would have been destroyed by fungal invaders now stand. They have started rewriting the chapters of the mosquito genome in ways that might, if we are careful and wise, end the oldest mass killer of children in history.
But the new authors also face questions that no author in history has faced. When you edit a somatic cell, you are revising one copy of the Book for one reader — the patient. When you edit the germline, you are revising the master edition from which all future copies will be printed. When you release a gene drive, you are rewriting the chapter for an entire species, in a library that you cannot fully control. These are different kinds of authorship. They carry different kinds of responsibility.
There is also the question of who gets to be an author. Right now, the people who can edit the Book are concentrated in wealthy countries and well-funded institutions. The readers who most need the revised editions — the children with sickle cell disease in Nigeria, the farmers watching their wheat crops collapse in Bangladesh — are often the ones with least access to the corrections that have been written for their specific chapters. This is the equity problem, and it is not a technical problem. It is a question of political will and institutional fairness that no CRISPR guide RNA can solve.
We began this series with Yoshizumi Ishino noticing, in 1987, a strange footnote in a bacterial genetics paper. Mysterious repeating sequences. He did not know what they meant. Francisco Mojica, a decade later, stared at similar sequences in the salt marshes of Alicante and would not let go of his obsession with them until he understood. Jennifer Doudna and Emmanuelle Charpentier sat together in a courtyard in Puerto Rico in 2011 and began a collaboration that would earn them the Nobel Prize and change medicine.
The Book is still being written. The question that this final cluster asks — and that no single cluster can answer — is not a scientific question. It is the oldest question in the human story: when we have the power to do something, how do we decide whether we should? The four-billion-year author left us no instructions. We are, for the first time, writing the Book ourselves. What we write will depend on what we value, what we fear, and how well we listen to each other. That is not a limitation of the science. It is the full measure of what the science has given us.
References & Further Reading
- Lander et al. (2019) — Adopt a moratorium on heritable genome editing. Nature 567:165. — A call for a global moratorium, signed by leading CRISPR scientists including pioneers of the field.
- Baltimore et al. (2018) — On human gene editing: International summit statement. Science 348:36. — The landmark statement from the first International Summit on Human Gene Editing.
- WHO Expert Advisory Committee (2021) — Human Genome Editing: Recommendations. WHO Press. — The WHO’s comprehensive governance framework for human genome editing.
- Esvelt & Gemmell (2017) — Conservation demands safe gene drive. PLOS Biology. — The ethical framework for responsible gene drive development.
- Doudna, J.A. — A Crack in Creation: Gene Editing and the Unthinkable Power to Control Evolution (2017). — Doudna’s own account of the CRISPR story, including her ethical reflections. Essential reading.
- Baylis, F. — Altered Genes, Twisted Truth: How Genetically Modified Organisms Have Corrupted Biotechnology — and her work specifically on consent and germline editing. francoisebaylis.ca
- Isaacson, W. — The Code Breaker: Jennifer Doudna, Gene Editing, and the Future of the Human Race (2021). — The most comprehensive narrative account of the CRISPR story and its ethical implications for a general audience.
- Somatic vs germline is the defining ethical line. Somatic editing: affects only the patient, not heritable, patient can consent — ethically acceptable with oversight. Germline editing: heritable, no consent possible from future persons — currently premature.
- He Jiankui crossed the line irresponsibly. The CCR5 experiment failed on every ethical standard: no medical necessity, known risks, incomplete editing, manipulated consent, no peer review. The scientific community’s condemnation was swift and appropriate.
- International governance is advisory, not binding. WHO framework, International Summit statements, and national laws provide guidance but cannot prevent a determined actor in a poorly regulated jurisdiction. Stronger international mechanisms are needed.
- Treatment vs enhancement is a genuine grey zone. The Alzheimer’s and deafness cases show that the boundary is philosophically complex. Clinical applications must be evaluated case by case, not resolved by a single principle.
- The consent problem for future persons is real and unresolved. Germline editing makes permanent decisions about people who cannot evaluate, consent to, or reverse those decisions. This is not merely procedural; it reflects a deep question about parental authority over children’s biological futures.
- Equity is not solved by good intentions. $2.2M price points, manufacturing complexity, and geographic inaccessibility mean the populations most affected by sickle cell disease have no access to the cure. Structural changes, not just rhetoric, are required.
- The 20-year horizon is extraordinary and uncertain. Ex vivo therapies will multiply, in vivo will expand, gene drives may be released, pig organs may become standard. The germline question will require a definitive answer by 2040. What that answer is will reflect the values — and the wisdom — of the societies that must make it.
Pingback: CRISPR in Agriculture & Animals