Professor Ian Brighthope has put forward an argument that deserves to be taken far more seriously than the scientific establishment is likely to take it. His proposition is deliberately radical: humanity should prohibit the deliberate genetic manipulation of viruses, bacteria and other self-replicating organisms because nobody has the moral authority to impose potentially irreversible biological consequences upon everyone else. His argument is not merely that gain-of-function research needs better committees, tighter laboratory procedures or another layer of regulation. It is that we have mistaken technological ability for moral permission.
Brighthope begins from a simple fact that becomes more unsettling the longer one thinks about it. Humanity inherited the biosphere; we did not design it. Yet small groups of researchers, corporations, governments, military establishments and private foundations increasingly possess the ability to alter organisms in ways that can reproduce, mutate and potentially spread beyond the laboratory. The people exposed to that risk have never consented to it, future generations obviously cannot consent to it, and other nations may have no practical means of preventing an engineered organism from crossing their borders. Brighthope therefore asks the question normally buried beneath technical discussions of biosafety: who gave anyone the right to do this?
The importance of the question lies in the distinction between an ordinary technological accident and a biological one. If an experimental aircraft crashes, the wreckage does not manufacture copies of itself. If a computer program fails, it does not ordinarily reproduce independently in forests, waterways, animals and human beings. A genetically altered replication-competent organism is different. Once released into a suitable environment, reproduction becomes part of the accident. Mutation and natural selection can then become part of it as well. The engineer may create the initial organism but cannot guarantee control over all its descendants.
This is where Brighthope attacks the seductive language of genetic "editing" and biological "programming." Such terminology encourages us to imagine DNA as software and the organism as a machine obediently executing instructions. But biological systems are vastly more complicated. Genes interact with other genes, regulatory systems, proteins, microorganisms and environmental conditions. A genetic intervention that behaves predictably under laboratory conditions does not thereby come with a guarantee covering thousands of generations in changing ecological circumstances. Evolution continues after the experimenter has published the paper and gone home.
The history Brighthope assembles demonstrates why the concern cannot simply be dismissed as science fiction. Researchers have synthesised infectious poliovirus from sequence information, constructed horsepox virus through synthetic biology, altered H5N1 influenza in experiments that produced airborne transmission between ferrets, and more recently demonstrated that generative AI can participate in designing viable bacteriophage genomes. Brighthope sees a technological progression here. We moved from modifying existing organisms to reconstructing pathogens from digital information and are now approaching a world in which computational systems can assist in designing biological entities that previously did not exist.
None of this establishes that catastrophe is inevitable. That would be the wrong argument. It establishes instead that the technological barrier is falling while the consequences of error remain potentially enormous. Knowledge cannot be recalled in the way that a defective consumer product can be recalled. Once techniques, sequences and procedures have been published and distributed, securing the original laboratory sample does not make the capability disappear.
Nor is laboratory infallibility a credible answer. Brighthope points to the embarrassing series of incidents disclosed by the US Centers for Disease Control and Prevention in 2014. The CDC acknowledged an incident involving potential exposure to viable anthrax and separately disclosed that a non-pathogenic avian influenza culture had been contaminated with highly pathogenic H5N1 and then shipped to another laboratory. These were not amateur operations conducted in somebody's garage. They occurred within one of the world's premier public-health institutions. The CDC responded with tighter safety measures, as it should have, but the larger philosophical point remains: human institutions cannot plausibly promise a zero-error future.
That becomes especially important when risk is asymmetric. Ten thousand experiments conducted without disaster do not prove that the ten-thousand-and-first cannot produce one. With technologies capable of creating self-replicating hazards, the calculation is unlike ordinary industrial risk because a sufficiently serious failure could be transboundary, persistent and perhaps irreversible. The relevant question therefore cannot merely be whether the probability of catastrophe appears small. We must also ask whether anyone is entitled to impose even that probability upon billions of people who receive none of the benefits and have no voice in the decision.
Gene drives make Brighthope's philosophical point particularly clear. A gene drive is intended to bias inheritance so that a selected genetic characteristic spreads through a population more efficiently than it would under normal Mendelian inheritance. The technology has potentially valuable applications, including controlling disease-carrying insects. Yet the very feature that makes a gene drive attractive, its capacity to propagate, is also what makes it ethically troublesome. The US National Academies has explicitly recognised concerns about potentially irreversible ecological consequences after release and concluded that proof-of-concept research was insufficient to justify environmental release.
Imagine that an engineered insect crosses a border. Which country consented on behalf of the species? Imagine that engineered genes enter wild populations and spread farther than predicted. Who recalls them? Imagine that suppressing one species changes the ecological niche occupied by another. Who calculated every subsequent interaction? These are not arguments that disaster must occur. They demonstrate something more fundamental: the scale of the intervention can exceed the political jurisdiction of the institution authorising it.
Brighthope's argument becomes still more disturbing when synthetic biology approaches forms of life without close natural precedents. He draws particular attention to proposals surrounding "mirror life," organisms constructed using mirror-image versions of biological molecules. Scientists themselves have warned about the potentially extraordinary hazards posed by self-replicating mirror organisms, precisely because normal biological recognition mechanisms might not function against them. Brighthope's response is that waiting until such organisms can actually be constructed before deciding that they should not be constructed would be extraordinarily foolish. By then the expertise, equipment and intermediate technologies required to create them may already be dispersed.
There is another dimension to the problem that has become impossible to ignore: dual use. Biotechnology cannot neatly separate knowledge useful for medicine from knowledge useful for biological attack. Techniques capable of changing host range, immune response, environmental persistence or transmissibility may have legitimate scientific purposes while simultaneously providing knowledge relevant to producing more dangerous organisms. The World Health Organization openly recognises this problem. Its global framework for responsible life-sciences research is explicitly concerned with accidental and deliberate misuse and seeks to manage dual-use risks while retaining the benefits of scientific innovation.
Brighthope parts company with that regulatory philosophy. WHO essentially asks how dangerous research can be governed responsibly. Brighthope asks whether certain categories of research should exist at all. That is the crucial divide. The regulatory approach assumes that benefits and risks can be balanced and that sufficiently elaborate governance can reduce the residual danger to an acceptable level. Brighthope argues that where the possible outcome involves an uncontrollable self-replicating biological agent, there may be no morally acceptable residual risk.
His position is therefore an application of the precautionary principle in its strongest form. Normally the precautionary principle has become a rhetorical ornament: governments invoke it when convenient and discover flexibility when economically or politically powerful interests want a technology deployed. Brighthope insists that genuine precaution has consequences. If an intervention could produce irreversible harm, uncertainty cannot automatically be counted in favour of proceeding. The fact that we do not know that catastrophe will occur is very different from knowing that it cannot occur.
There is considerable force in that distinction. Conventional risk analysis works reasonably well where consequences are bounded and reversible. We accept bridges, aircraft, pharmaceuticals and power stations despite their risks because experience permits probabilities to be estimated and failures usually remain limited in space and time. A technology capable of producing an indefinitely reproducing agent poses a different problem. If the worst outcome includes sustained transmission or permanent ecological establishment, then conventional calculations involving expected benefits and estimated probabilities begin to look philosophically inadequate.
Brighthope also identifies the institutional problem behind such calculations. The benefits of proceeding are concentrated. Researchers obtain publications and careers, universities receive grants and intellectual property, biotechnology corporations acquire products, governments gain strategic capabilities and funding organisations demonstrate results. The potential costs, meanwhile, are dispersed across people who never entered the laboratory and perhaps across generations not yet born. This does not require scientists to be corrupt. It merely requires them to be human. Institutions naturally become enthusiastic about activities from which their own institutions benefit.
The standard response is that genetic engineering has also produced enormous benefits, and Brighthope does not pretend otherwise. Recombinant technologies have contributed to insulin, clotting factors, vaccines, diagnostic materials and other valuable products. An intelligent critique cannot simply wave those achievements away. Brighthope therefore proposes transitional exemptions for established production systems supplying essential medicines while alternatives are developed. His target is the continuing creation and alteration of replication-competent organisms, particularly where the resulting organisms themselves constitute the experiment or are intended for environmental propagation.
Whether one follows him all the way to a comprehensive prohibition is therefore a separate question from whether his underlying argument succeeds. There are legitimate objections to an absolute ban. Genetic modification can potentially reduce disease, improve agriculture and produce medicines, while an overly broad prohibition could suppress relatively low-risk research along with genuinely dangerous work. Enforcement would also be formidable. A treaty observed by democracies but ignored by secret military programs or rogue states could conceivably leave responsible countries scientifically disadvantaged without eliminating the danger.
Yet that objection cuts both ways. If the technology is dangerous enough that nations believe they must continue developing it because other nations might do so, we have recreated the logic of an arms race. "We cannot stop because somebody else might continue" is precisely the reasoning that produces escalating capabilities nobody particularly wants but everybody fears being without.
Brighthope therefore proposes something much more ambitious than improved laboratory regulation: an international convention prohibiting specified forms of manipulation of replication-competent life, backed by national legislation, inspections, funding transparency, whistleblower protection and liability for institutions and individuals. The prohibition would extend to military and private programs rather than allowing "national security" or philanthropic funding to become escape clauses. Observation and sequencing of naturally occurring organisms could continue, as could genuinely non-replicating and cell-free research. The boundary would be drawn around the deliberate creation of inheritable biological novelty capable of propagating beyond its creator's continuing control.
There would undoubtedly be difficult boundary cases, and Brighthope's proposal would require far more detailed legal drafting before it could become a workable treaty. But difficult boundaries do not destroy the principle. International law already accepts that some technically possible activities should be prohibited because their consequences are intolerable. Biological weapons provide the obvious precedent. Humanity did not conclude that every conceivable military use of pathogenic organisms had to be assessed individually by a committee. It decided that an entire category of activity crossed a line.
The deeper value of Brighthope's argument is consequently philosophical rather than merely technical. Modern technological civilisation tends to assume that whatever can be done eventually will be done and that the only serious question is how it should be regulated. That assumption deserves rejection. Human beings can acquire powers that they are not wise enough to exercise. Scientific knowledge expands our ability to intervene in nature much faster than it expands our ability to foresee every consequence of intervention.
For most of human history our mistakes were constrained by our weakness. We could damage a valley, poison a river or destroy a local population, but we lacked the ability to redesign mechanisms of inheritance and send those changes reproducing through the biological world. Biotechnology is progressively removing that constraint. AI and automated biological design may accelerate the process further. Our technical power is becoming global while our foresight remains stubbornly human.
That is why Brighthope's central question cannot be answered merely by another assurance that laboratories have protocols. The question comes before the protocol. By what moral authority does a researcher, corporation, foundation or government accept an irreversible biological risk on behalf of humanity?
There may be cases in which society ultimately decides that the benefit justifies carefully circumscribed genetic intervention. But Brighthope is right that the presumption deserves to be reversed when self-replication and potentially irreversible spread are involved. The advocates of the experiment should bear the burden of demonstrating why everyone else must accept the risk. Scientific curiosity, commercial advantage and technological prestige are not sufficient answers.
The most dangerous idea in this debate may therefore be the apparently moderate one: that progress is inevitable and regulation will somehow catch up. Regulation can close a laboratory. It can withdraw a grant, punish a researcher or prohibit another experiment. What it cannot necessarily do is retrieve a self-replicating organism after that organism has escaped into the evolutionary world.
Civilisation has learned, usually after painful experience, that restraint is itself a technological achievement. Wisdom consists not only in discovering how to open doors but in recognising doors that should remain closed. Brighthope's challenge is to decide where that boundary lies before an accident decides it for us.
The biosphere is not laboratory property. It is the common inheritance of humanity and of generations that have not yet been born. Those who wish to rewrite parts of it therefore carry a moral burden vastly greater than the possession of a grant, a biosafety certificate or the technical ability to perform the experiment.
Being able to rewrite life is an extraordinary scientific achievement. Knowing when not to do so may prove an even greater one.
https://ianbrighthope.substack.com/p/no-one-has-the-right-to-rewrite-the