Synthetic Biology — The Revolution Nobody Is Governing

To grasp the scale of what is under way, begin with the investment. The synthetic biology sector attracted an estimated $17 billion in venture capital and public funding by the end of 2025 [1]. The CRISPR gene editing market alone — one segment within the broader field — was valued at $4.46 billion in 2025, projected to reach $14.96 billion by 2035 [1]. Genome engineering accounts for 33.21% of the total synthetic biology market, with the Asia-Pacific region growing at 15.18% annually — faster than any other region on earth [1]. These are not speculative projections. They are observations about capital already deployed.

The investment follows the science. In December 2023, the US Food and Drug Administration approved Casgevy — the first CRISPR-based gene therapy ever authorised — for the treatment of sickle cell disease [4]. In March 2024, Massachusetts General Hospital performed the first transplant of a gene-edited pig kidney into a living human [9]. By the end of 2024, China had approved five gene-edited crop varieties for cultivation [13]. In 2025, a research team in Tanzania demonstrated that gene-drive-capable mosquitoes could suppress real-world malaria in contained laboratory conditions [2]. Each of these milestones, individually, would have been science fiction two decades ago. Together, they describe a technological revolution advancing faster than any governance framework can track.

The breadth of application is what distinguishes synthetic biology from earlier biotechnology waves. This is not a single-purpose technology. Gene editing is simultaneously being deployed in human therapeutics (Casgevy for sickle cell, over 100 clinical trials worldwide), agriculture (drought-resistant maize in Brazil, high-yield wheat in China), xenotransplantation (pig organs with up to 69 gene edits for human recipients), and ecological intervention (gene drives designed to suppress malaria-carrying mosquito populations) [1] [13]. The same underlying CRISPR-Cas9 technology that cures a genetic blood disorder in a hospital in Nashville is being used to redesign an entire species of mosquito in a laboratory in Tanzania. The distance between those two applications — one individual, one ecological; one reversible, one potentially permanent — defines the governance challenge.

What makes the current moment unprecedented is not merely the power of the technology but the speed of its dissemination. In 2012, Jennifer Doudna and Emmanuelle Charpentier published the foundational paper demonstrating CRISPR-Cas9 as a programmable gene editing tool. Twelve years later, it is routine in over 50 countries [15]. The cost of gene synthesis has fallen by a factor of 1,000 since 2000. A university laboratory can now perform gene editing for a few hundred dollars using commercially available kits. The democratisation of the technology is, depending on one's perspective, either the greatest opportunity in the history of medicine or the most dangerous development in the history of biosecurity. The evidence suggests it is both.

The question this report examines is not whether synthetic biology is transformative — that is no longer in dispute. The question is whether governance is keeping pace with a technology that can rewrite the genetic code of any living organism, create organisms that have never existed in nature, and — through gene drives — propagate genetic modifications through entire wild populations with no demonstrated mechanism for recall. The evidence, as the following sections will show, is that it is not.

Standard genetic inheritance follows Mendelian rules: a modified gene has a 50% chance of being passed to each offspring. Gene drives circumvent this constraint. By encoding the CRISPR machinery alongside the desired modification, a gene drive copies itself onto both chromosomes during reproduction, achieving inheritance rates approaching 100% [2]. In theory, releasing a small number of gene-drive-modified organisms into a wild population could propagate a genetic change through the entire species within a handful of generations. This is not theoretical speculation — it has been demonstrated in contained laboratory populations of mosquitoes, fruit flies, and mice.

Two principal gene drive strategies have emerged. Suppression drives aim to reduce or eliminate a target population entirely — for example, by spreading a gene that causes female infertility in Anopheles gambiae mosquitoes, driving the species toward local extinction. Modification drives take a different approach: rather than eliminating the species, they alter it — for instance, by spreading a gene that makes mosquitoes unable to carry the Plasmodium parasite that causes malaria [2]. The Tanzania study published in Nature in 2025 demonstrated a modification approach: engineering Anopheles gambiae to robustly inhibit genetically diverse strains of Plasmodium falciparum obtained from naturally infected children [2].

The distinction between these approaches matters enormously for risk assessment. A modification drive that fails simply leaves mosquitoes unchanged — they continue to exist but retain their capacity to transmit malaria. A suppression drive that succeeds removes an entire species from an ecosystem, with cascading consequences for every organism that feeds on, competes with, or is pollinated by that species. A suppression drive that partially succeeds may be worse: creating evolutionary selection pressure that drives resistance, potentially producing mosquito populations that are harder to target with any subsequent intervention [12].

Beyond mosquitoes, gene drives are being researched for invasive species control — eliminating invasive rodents from island ecosystems, suppressing agricultural pests such as the spotted-wing drosophila, and even modifying tick populations to reduce Lyme disease transmission. Each application carries the same fundamental characteristic: once a self-propagating gene drive is released into a wild population, there is currently no proven mechanism to reverse it [12]. The French Academy of Sciences stated plainly in September 2025 that the technique "presents a variety of potential dangers and is uncontrollable" [12].

The parallel development of AI-powered biological design tools has accelerated this trajectory. Generative protein design platforms — most notably RFDiffusion, developed in the laboratory of 2024 Nobel laureate David Baker — can now design novel proteins with specified functions [7]. The convergence of AI and synthetic biology — what researchers term SynBioAI — means that the design cycle for new biological constructs is compressing from years to months to weeks. The OECD convened 66 experts from six continents in 2023-2024 specifically to assess this convergence, concluding that governance frameworks must urgently adapt [5].

The technology's supporters argue that so-called "daisy chain" gene drives — engineered to lose potency after a fixed number of generations — could provide a self-limiting mechanism. Theoretical models suggest such drives would remain geographically contained and temporally limited. But no daisy chain drive has been tested in wild populations, and the gap between theoretical models and ecological reality is precisely where the risk resides. Evolution is relentlessly creative in finding ways to circumvent engineered constraints — a lesson that antibiotic resistance, pesticide resistance, and herbicide resistance have taught repeatedly.

The approval of Casgevy in December 2023 represented a watershed. The therapy uses CRISPR-Cas9 to edit a patient's own haematopoietic stem cells, reactivating foetal haemoglobin production to compensate for the defective adult haemoglobin that causes sickle cell disease. In clinical trials, patients who received the one-time treatment were free of severe vaso-occlusive crises — the excruciating pain episodes that define the disease — for at least twelve months [4]. For a disease that affects approximately 100,000 Americans — disproportionately Black — and millions worldwide, this is transformative. But transformation at $2.2 million per treatment raises its own questions.

The economics of Casgevy illuminate the structural tensions within gene therapy. With approximately 16,000 eligible US patients, the total addressable market exceeds $35 billion at list price [4]. The treatment requires a specialist transplant centre, myeloablative conditioning (chemotherapy to destroy the patient's existing bone marrow), and weeks of hospitalisation. In the United States, insurance coverage and patient assistance programmes partially mitigate the cost. In sub-Saharan Africa, where sickle cell disease is most prevalent, the treatment is effectively inaccessible. The technology exists to cure a genetic disease. The delivery system does not exist to cure it equitably.

Xenotransplantation — the transplantation of animal organs into humans — has advanced with equal speed and equal complexity. In March 2024, Massachusetts General Hospital performed the first transplant of a gene-edited pig kidney into a living human recipient [9]. By November 2024, Towana Looney, a 53-year-old woman from Alabama, became the third recipient of a gene-edited pig kidney at NYU Langone and was discharged eleven days after surgery [9]. The eGenesis pig kidney involves 69 separate gene edits — removing pig genes that trigger human immune rejection and adding human genes to improve compatibility — making it the most extensively engineered animal organ ever transplanted into a human [9].

The agricultural domain tells a parallel story of rapid deployment. China approved five gene-edited crop varieties by the end of 2024 — including soybean, wheat, corn, and rice varieties with enhanced yield and nutritional profiles — representing a major policy shift for a country that had previously maintained restrictive GMO regulations [13]. Brazil launched national field trials of CRISPR-edited drought-resistant maize varieties in October 2024, becoming the first country with a government-backed genome editing crop initiative [13]. Japan approved a gene-edited tomato with increased GABA content — marketed as a health food — making it among the first gene-edited foods available to consumers anywhere in the world [13].

In the United States, the regulatory pathway for gene-edited crops has been notably permissive. The USDA does not regulate gene-edited plants that could have been produced through conventional breeding — a category that encompasses most CRISPR-edited crops with single-gene deletions or modifications. Norfolk Healthy Produce developed a purple tomato with enhanced antioxidant properties; GreenVenus created non-browning lettuce and avocado varieties [13]. Industry analysts project that 2026 could be the year when whole CRISPR-edited fruits and vegetables reach mainstream supermarkets. The EU, by contrast, has maintained its precautionary approach — though regulatory reform proposals were under consideration through 2025.

The pattern across all three domains — therapeutics, xenotransplantation, agriculture — is consistent: rapid scientific progress, accelerating commercialisation, enormous potential benefit, and governance frameworks that vary wildly between jurisdictions and struggle to keep pace with the technology they are meant to regulate. What is approved in Japan may be prohibited in the EU. What costs $2.2 million in Nashville is unavailable in Lagos. The technology is global. The governance is not.

The numbers are staggering in their persistence. In 2024, the WHO recorded 282 million malaria cases worldwide, up from previous years despite decades of intervention [10]. Three countries alone — Nigeria (31.9%), the Democratic Republic of the Congo (11.7%), and Niger (6.1%) — accounted for over half of all malaria deaths globally [10]. Conventional interventions — bed nets, indoor spraying, artemisinin-based combination therapies — have averted an estimated 2.3 billion cases and 14 million deaths since 2000. But progress has stalled, and antimalarial drug resistance has been confirmed or suspected in eight countries [10]. The parasite is adapting faster than the treatments.

It was into this context that Target Malaria — a not-for-profit research consortium funded by the Bill & Melinda Gates Foundation and led by Imperial College London — pursued the most ambitious gene drive programme in the world. The goal: engineer Anopheles gambiae mosquitoes with a gene drive that would spread female infertility through wild populations, suppressing the primary malaria vector species in sub-Saharan Africa [3].

On August 11, 2025, Target Malaria released 75,000 second-phase genetically modified mosquitoes in Burkina Faso [3]. These were not gene drive mosquitoes — they were an intermediate step, modified male mosquitoes carrying a marker gene, intended to demonstrate the capacity for controlled release and monitoring before any gene drive deployment. Eleven days later, on August 22, the government of Burkina Faso announced the suspension of all Target Malaria activities throughout the entire national territory [3]. Facilities housing genetically modified mosquitoes were sealed. All samples were destroyed.

The reasons for the suspension were multiple and contested. Critics — including environmental organisations, indigenous rights groups, and some biosafety researchers — had long argued that Target Malaria was advancing too quickly, that community consent processes were inadequate, and that the ecological risks of gene drive deployment had not been sufficiently characterised [12]. Andrea Crisanti, the Imperial College London researcher who heads the lead gene drive laboratory, subsequently acknowledged that the proposed gene drive strain had "significant flaws" with "multiple implications for disease transmission and ecological adaptation" [3].

The Burkina Faso episode crystallises the central tension in gene drive governance. On one side: 610,000 malaria deaths per year, rising drug resistance, and a technology that could — in principle — dramatically reduce transmission by modifying the mosquito vector. On the other: a technology whose environmental consequences are irreversible, whose ecological effects are poorly understood, whose governance framework is non-existent at the international level, and whose most ambitious trial was suspended by the host government within eleven days of its initiation. Neither the urgency of the disease burden nor the precaution demanded by irreversibility can be wished away. Both are simultaneously valid.

Meanwhile, the science continues to advance. The Tanzania study published in Nature in 2025 demonstrated that gene-drive-capable mosquitoes — engineered locally in Tanzanian facilities — could robustly inhibit genetically diverse Plasmodium falciparum isolates from naturally infected children [2]. Unlike Target Malaria's suppression approach, this modification approach does not aim to eliminate mosquitoes — it aims to make them unable to carry the parasite. If successfully deployed, it would leave mosquito populations intact while eliminating their role as malaria vectors. The distinction is significant: modification drives pose lower ecological risk than suppression drives, because they do not aim to remove a species from an ecosystem.

The most alarming development involves the intersection of AI protein design tools and DNA synthesis technology. In 2025, a team of Microsoft researchers demonstrated that open-source AI tools could be used to engineer new protein variants of known pathogens that successfully evaded existing DNA synthesis screening procedures [7]. This is not a theoretical vulnerability. It is a demonstrated capability. The screening systems that DNA synthesis companies use to prevent customers from ordering dangerous sequences were circumvented by AI-designed modifications that preserved the protein's function while altering its sequence enough to avoid detection [7].

The DNA synthesis screening system itself is fragile. The Biden administration introduced a Framework for Nucleic Acid Synthesis Screening in 2024 that guides manufacturers of bench-top synthesis equipment to screen purchase orders for sequences of concern and assess customer legitimacy [11]. But a May 2025 White House Executive Order created uncertainty over the status of this framework, and the Arms Control Association warns that "regulatory gaps in bench-top nucleic acid synthesis create biosecurity vulnerabilities" [7]. The increasing availability of desktop DNA synthesisers — machines that can produce custom DNA sequences without relying on commercial synthesis providers — further undermines any screening regime based on provider-side checks.

The democratisation of the technology extends beyond professional laboratories. In the United States alone, over 50 community biology laboratories (DIYbio spaces) operate with nearly 30,000 participants — many without formal biosafety training [6]. Globally, the DIYbio movement includes approximately 60 groups in Europe, 22 in Asia, 16 in Latin America, and a growing number in Africa. The FBI's Weapons of Mass Destruction Directorate has engaged with this community, and most assessments find it largely safety-conscious and self-regulating. But the infrastructure for gene editing at home or in community spaces is no longer speculative — it is commercially available, affordable, and increasingly capable.

The National Academies of Sciences, Engineering, and Medicine issued a report in March 2025 assessing AI capabilities in biological design. The report concluded that current state-of-the-art AI tools can design relatively simple biological structures such as individual proteins and molecules, but are currently unable to design self-replicating pathogens — and it is unlikely that currently available viral sequence data are sufficient to train such a model [14]. This assessment provides some reassurance about the current threat level. But the operative word is "current" — AI capabilities in biological design are advancing rapidly, and the same report noted that the gap between current capabilities and dangerous ones is narrowing.

At the Munich Security Conference in February 2024, NTI launched the International Biosecurity and Biosafety Initiative for Science (IBBIS) — an independent organisation headquartered in Geneva dedicated to reducing risks from bioscience research, with an initial focus on preventing the misuse of DNA synthesis technology [14]. IBBIS represents an institutional acknowledgement that the existing governance architecture — the Biological Weapons Convention, the Cartagena Protocol, national biosafety regulations — was not designed for an era in which the tools of pathogen creation are becoming cheaper, more accessible, and increasingly AI-enhanced.

The Carnegie Endowment for International Peace articulated the governance deficit bluntly in October 2024: "Current national and international regulatory mechanisms do not adequately address the rising biosecurity risks that accompany this development" [6]. The report identified a specific structural problem: biosecurity governance is distributed across multiple treaties, agencies, and national frameworks, none of which has clear jurisdiction over the convergence of AI and synthetic biology. This is not regulatory underperformance. It is a governance architecture built for a different technological era.

The biosecurity challenge is qualitatively different from the ecological and medical governance challenges posed by gene editing. Gene drives threaten irreversible ecological change. Gene therapies raise questions of equitable access. But biosecurity failures could be catastrophic on a different scale — the deliberate or accidental creation of a novel pathogen using tools that are increasingly available, affordable, and powerful. The OECD concluded that the convergence of synthetic biology, AI, and automation requires governance frameworks that do not yet exist — and that building them is not merely advisable but urgent [5].

The country-by-country regulatory comparison reveals deep structural fragmentation. The United States regulates biotechnology products through three separate agencies — the USDA, the FDA, and the EPA — under a Coordinated Framework for the Regulation of Biotechnology established in 1986, decades before CRISPR existed. Gene-edited crops that could have been produced through conventional breeding are largely exempt from USDA regulation [11]. Japan adopted a similar permissive approach in 2020, allowing gene-edited foods to be sold without safety evaluations provided the modifications meet certain criteria [13].

The European Union has historically taken a precautionary approach, subjecting gene-edited organisms to the same rigorous approval process as conventional GMOs under Directive 2001/18/EC. Reform proposals circulated through 2024-2025 would create a tiered system — lighter regulation for simple edits that mimic natural mutations, stricter oversight for more complex modifications — but legislative progress has been slow [15]. The UK, post-Brexit, has moved in the opposite direction: establishing the Regulatory Innovation Office in October 2024 with the explicit aim of "streamlining regulatory procedures" for synthetic biology products [15].

China presents perhaps the most complex regulatory landscape. Following the He Jiankui scandal, China enacted the 11th Amendment to its Criminal Law in December 2020, specifically criminalising heritable human genome editing — the first country to establish a criminal penalty for the practice [8]. Yet Chinese law generally permits somatic cell gene editing research, and China has simultaneously become one of the most permissive jurisdictions for agricultural gene editing, approving five crop varieties in 2024 [13]. Analysts have identified "fragmented oversight, an unclear allocation of legal responsibilities, and insufficient capacity within ethical review committees" as persistent structural weaknesses [8].

At the international level, the architecture is thinner still. The Cartagena Protocol on Biosafety — adopted in 2000, entered into force in 2003 — was designed for an era of conventional GMOs and does not specifically address gene drives, synthetic biology, or AI-designed organisms [15]. The Convention on Biological Diversity has prepared "voluntary guidance materials for risk assessments of living modified organisms containing engineered gene drives" — but the operative word is voluntary. The Biological Weapons Convention prohibits the development of biological weapons but has no verification mechanism and no enforcement capacity. The Kunming-Montreal Global Biodiversity Framework, adopted by 196 countries in December 2022, includes a biosafety target but relies on voluntary compliance [15].

The He Jiankui affair illustrates both the consequences of governance failure and the limits of reactive regulation. In November 2018, He announced the birth of the world's first gene-edited babies — twins named Nana and Lulu, whose CCR5 gene he had attempted to modify using CRISPR to confer HIV resistance [8]. The experiment violated Chinese regulations, involved forged ethics approval documents, and breached international norms on informed consent. He was sentenced to three years in prison and fined three million yuan ($429,000) [8]. Released in April 2022, he opened a new laboratory in Beijing in November 2022, shifting his focus to less controversial somatic gene therapies for rare diseases [8].

The He Jiankui case prompted China's criminal law amendment — governance by scandal. The Burkina Faso suspension was governance by crisis. The pattern is consistent: governance responds to events rather than anticipating them. For a technology whose defining characteristic is irreversibility, reactive governance is structurally inadequate. The gene drive released today cannot be un-released after tomorrow's investigation concludes it was premature.

The comparison is not symmetrical, and responsible analysis must acknowledge this. The costs of excessive caution are measurable in deaths — 610,000 malaria deaths per year, thousands dying on organ transplant waiting lists, patients suffering from curable genetic diseases. The costs of insufficient caution are speculative but potentially catastrophic — an uncontrollable gene drive spreading through ecosystems, an AI-designed pathogen released deliberately or accidentally, a cascade of ecological effects from species suppression that cannot be reversed.

The precautionary principle — the idea that technologies should be proven safe before deployment — has been the dominant framework in European and international biosafety governance. But the principle struggles when the alternative to deployment is also deadly. A strict application of precaution to gene drive technology means accepting 610,000 malaria deaths per year while waiting for certainty about ecological risks that may never be fully resolved. This is not a comfortable position. It is, however, a position that takes irreversibility seriously — and the history of technological intervention in complex systems suggests that irreversibility deserves serious weight [12].

The proactionary principle — the idea that the burden of proof should fall on those who would restrict innovation rather than those who would deploy it — has gained influence particularly in the United States and the United Kingdom. Biden's Executive Order 14081 explicitly aims to "advance biotechnology and biomanufacturing innovation" with a whole-of-government approach [11]. The UK's Regulatory Innovation Office explicitly aims to "streamline regulatory procedures" [15]. Both framings assume that the default risk of regulation is excessive caution. Neither framework adequately addresses the specific challenge of irreversibility.

The intellectual honesty required by this debate demands acknowledging that both sides are partially right — and that the resolution is not a choice between acceleration and restraint but a redesign of governance that can manage both simultaneously. The current frameworks — precautionary in Europe, proactionary in the US and UK, reactive everywhere — are not adequate for a technology that is simultaneously essential (malaria), transformative (gene therapy), and potentially catastrophic (biosecurity, ecological irreversibility).

The structural deficits are now well-documented. The Carnegie Endowment has identified the governance gap [6]. The OECD has mapped the convergence risks [5]. The Arms Control Association has documented the biosecurity vulnerabilities [7]. The French Academy of Sciences has declared gene drives "uncontrollable" [12]. IBBIS has been established in Geneva precisely because existing institutions are inadequate [14]. The evidence base for the governance deficit is no longer contested by any credible institution. The question is whether the political will exists to act on it.

Several structural principles emerge from the evidence. First, governance must be anticipatory rather than reactive. The pattern of governance-by-scandal (He Jiankui) and governance-by-crisis (Burkina Faso) is incompatible with technologies whose consequences are irreversible. If a gene drive causes ecological damage, the investigation that follows cannot undo it. If an AI-designed pathogen escapes containment, the regulatory response cannot recall it. Anticipatory governance requires the capacity to evaluate technologies before deployment — which in turn requires international coordination, scientific expertise within regulatory bodies, and decision-making frameworks that can operate at the speed of the technology [5].

Second, international coordination is not optional — it is a structural necessity. Synthetic biology does not respect national borders. A gene drive released in Burkina Faso can spread to Mali, Niger, Ghana, and beyond. A pathogen engineered in one country can infect populations in every other. A gene-edited crop approved in China can cross-pollinate with wild relatives in any neighbouring ecosystem. The current framework of voluntary guidelines and national regulations is structurally incapable of managing transboundary biological risks [15]. What is needed is not another voluntary framework but a binding international instrument — comparable in ambition, if not in specifics, to the Nuclear Non-Proliferation Treaty — that establishes minimum standards for gene drive governance, biosecurity screening, and the regulation of heritable genome editing.

Third, equity must be a design principle, not an afterthought. The current trajectory of synthetic biology is producing a two-tier world: wealthy nations developing gene therapies at $2.2 million per patient and gene-edited crops for commercial agriculture, while the populations who would benefit most from gene drive malaria suppression — children in sub-Saharan Africa — have no meaningful voice in governance decisions that will shape their future. The OECD's 66-expert consultation drew from six continents [5]. That level of geographic and economic diversity must be structural — embedded in governance institutions, not added as consultation theatre.

Fourth, the convergence of AI and synthetic biology requires governance mechanisms that do not yet exist anywhere. Current biosafety and biosecurity frameworks were designed for an era in which biological modification required specialised expertise and expensive equipment. The era of AI-designed proteins, desktop DNA synthesisers, and community biology laboratories demands a different model — one that addresses the design tools themselves, not just the organisms they produce. The OECD's planned Recommendation for Responsible Innovation in Synthetic Biology is a step in this direction, but recommendations are not regulations [5].

The evidence presented in this report describes a technology of extraordinary promise deployed in a governance vacuum. CRISPR can cure genetic diseases. Gene drives could save hundreds of thousands of lives from malaria. Gene-edited crops could enhance food security under climate stress. Xenotransplantation could resolve the organ shortage crisis. These are not speculative benefits — they are demonstrated capabilities, being deployed now, in hospitals and fields and laboratories across the world.

But the same technology enables the creation of heritable human genetic modifications by a single researcher who forges ethics documents. It enables the release of organisms into ecosystems from which they cannot be retrieved. It enables the AI-assisted design of proteins that evade the biosecurity screening systems meant to prevent weaponisation. And it is governed by a patchwork of national regulations, voluntary international frameworks, and institutions designed for a different technological era.

The question is not whether synthetic biology will reshape the world. It already is. The question is whether humanity will build the governance architecture to manage that reshaping before an irreversible mistake forecloses the options available to future generations. The $19 billion market will not wait. The gene drives will not wait. The 610,000 malaria deaths per year will not wait. The only variable still within human control is whether governance can match the pace of the revolution it is meant to govern. The evidence, as of 2026, suggests it cannot — not yet.