The biotechnology revolution has moved at a pace that regulatory systems were never designed to match. Gene editing with CRISPR has gone from laboratory curiosity to clinical reality in less than a decade. Artificial intelligence can now design novel proteins with structural or functional properties similar to known hazardous ones — potentially evading detection systems. DNA synthesis has become so affordable and distributed that traditional gatekeeping mechanisms no longer work. And across these converging domains — synthetic biology, gene therapy, and stem cell manufacturing — national and international biosafety frameworks remain fragmented, inconsistently enforced, and in some cases, obsolete.
A comprehensive review published in the Journal of Biosafety and Biosecurity catalogs where the gaps are most acute and why they matter for biosecurity. The picture that emerges is of a biotechnology landscape where the pace of innovation has decisively outrun the coordination of governance, creating exploitable vulnerabilities at the intersection of research, clinical translation, and potential misuse.
The Synthetic Biology Problem: Dual-Use Tools Becoming Universally Accessible
CRISPR-based genome editing exemplifies the governance challenge. The technology has become faster, cheaper, and more accessible — developments with obvious therapeutic benefits and equally obvious misuse potential. But oversight of the DNA synthesis supply chain, which is foundational to controlling access to dangerous genetic material, remains inconsistent. Provider screening varies significantly across vendors, and much of the guidance is voluntary and nonbinding.
The problem is compounded by artificial intelligence. Recent research has shown that AI systems trained on large protein and genomic datasets can generate novel amino acid sequences with structural or functional properties similar to known hazardous proteins — but different enough that they may not be flagged by the similarity-based detection tools that some DNA synthesis services rely on. If this capability were deployed at scale, it could potentially lower the time, expertise, and cost required to engineer dangerous biological phenotypes.
Beyond gene editing, the landscape includes other underappreciated dual-use domains that merit close attention. Industrial bioproduction platforms designed for legitimate metabolic engineering could potentially be redirected to manufacture harmful metabolites or biologically active toxins. Gene-drive technologies, engineered to spread traits through wild animal populations for vector control or ecological management, raise significant environmental risks if unintended releases occur. Cell-free synthetic biology systems, which enable biological reactions outside living cells, reduce traditional laboratory containment barriers and could facilitate decentralized production of biologically active molecules. Emerging xenobiology approaches employing non-canonical nucleic acids or expanded genetic codes create orthogonal biological systems whose environmental interactions and long-term biosafety implications remain incompletely understood.
Cybersecurity threats are also rising. Malicious access to cloud-based genomic databases and automated laboratory platforms could enable theft of pathogen data, tampering with experimental processes, or circumvention of DNA synthesis screening. Digital infrastructure in life sciences — increasingly dependent on cloud-based platforms, automated sequencing systems, and bioinformatics pipelines — represents a new frontier for biosecurity oversight that most regulatory frameworks do not yet address.
Gene Therapy: Known Risks Operating Without Adequate Monitoring
The clinical reality of gene therapy reveals biosafety hazards that have moved from theoretical to documented. CRISPR-based editing tools cause unintended off-target modifications at measurable rates: chromosome translocations, deletions, and in more severe cases, large-scale genomic rearrangements resembling chromothripsis. In some studies using sensitive sequencing techniques, structural abnormalities have been identified in approximately 15 to 20 percent of edited cells, though these findings vary across studies and detection methods.
Viral vectors — the dominant delivery mechanism for gene therapies — carry their own profile of risks. Adeno-associated virus (AAV) vectors have been associated with immune-mediated hepatotoxicity at high systemic doses, a serious concern for liver-directed therapies. Lentiviral vectors, which integrate into the host genome, carry insertional mutagenesis risk — the possibility that random integration near oncogenes could trigger carcinogenic pathways.
Following systemic administration, transient vector DNA has been detected in blood, urine, saliva, and semen — findings that underscore the need for environmental surveillance and long-term follow-up protocols. Yet most jurisdictions do not yet mandate such protocols as standard practice. Cytokine release syndrome, well-documented in CAR-T cell therapies, represents an additional immunological hazard requiring strict clinical monitoring.
Beyond acute effects, gene therapy can produce long-term consequences. If regulatory or metabolic genes are affected, persistent dysregulation of cellular and metabolic pathways can occur. Some high-dose systemic gene therapy models have shown persistent low-level inflammation or altered tissue physiology. Although adverse outcomes from genomic integration events are rare, long-term monitoring remains necessary given the potential for dysregulated gene expression or malignant transformation, particularly in dividing tissues.
Yet globally, oversight of these complications remains fragmented. There is no internationally consistent registry tracking adverse events across gene therapy trials, no standardized reporting mechanism for off-target edits or long-term epigenetic effects, and no coordinated mechanism for detecting patterns of complications that individual trials might miss.
Stem Cell Manufacturing: Oncogenic Risk in an Unequal Regulatory Environment
Stem cell therapies introduce hazards arising from the manufacturing process itself. The same molecular pathways that allow pluripotent stem cells to self-renew and differentiate can, if dysregulated, drive malignant transformation. Residual undifferentiated cells can form teratomas following transplantation. Extended culture introduces genomic aberrations — copy number variations, aneuploidy, and chromosomal rearrangements that resemble cancer-associated changes. At manufacturing scale, risks include cross-contamination between cell lines, microbial pathogen introduction, and viral vector escape.
Addressing these risks requires strict cGMP compliance, comprehensive cell characterization, and coordinated international safety standards. Yet high-income countries maintain structured oversight while many low- and middle-income countries lack equivalent infrastructure, allowing unproven stem cell therapies to proliferate with no coordinated framework to bridge the divide.
A Fragmented Global Landscape
The regulatory architecture remains fundamentally disjointed. NIH guidelines require institutional biosafety committee review for recombinant and synthetic nucleic acid research but offer limited explicit coverage for AI-driven design workflows or complex synthetic biology constructs. WHO laboratory biosecurity guidance is risk-based and comprehensive but non-binding and lacks enforceable global surveillance mechanisms. The EU’s Advanced Therapy Medicinal Product regulation provides a centralized framework but is unevenly implemented across member states and offers insufficient guidance for engineered synthetic circuits or AI-designed modifications. FDA gene therapy guidance continues to evolve but remains internationally inconsistent. ISO standards for stem cell biobanking exist but are incompletely adopted, particularly in lower-income settings.
The result is a patchwork: researchers and companies operating in high-regulation jurisdictions face compliance costs that incentivize offshoring; jurisdictions with weaker oversight become attractive destinations for clinical trials and manufacturing; and emerging threats — AI-assisted biological design, cyberbiosecurity, novel dual-use platforms — fall through regulatory gaps because no single framework was designed to address them.
Scientific uncertainty compounds the governance challenge. Rapidly evolving technologies like CRISPR-based editing, synthetic gene circuits, and AI-assisted design introduce new uncertainties in risk prediction: off-target genomic effects, long-term cellular behavior, ecological impacts following organism release. These scientific unknowns complicate regulatory evaluation and biosafety risk assessment at precisely the moment when frameworks need to be most robust.
What Needs to Change
The solutions are clear but ambitious. An integrated, real-time global biosafety registry is needed to track adverse events across gene and cell therapy trials — capturing vector-related complications, off-target edits, immune toxicities, and long-term epigenetic effects. DURC governance frameworks must be explicitly redesigned to address AI-assisted biological design, incorporating computational sequence screening and export-control mechanisms. Multi-omics risk evaluation — combining transcriptomics, proteomics, metabolomics, and immunogenicity profiling — is necessary to capture complex off-target effects that single-endpoint assays miss.
Alignment of GMP standards, quality assurance protocols, and ethical accountability frameworks are prerequisites for equitable and safe clinical translation.
Capacity building biosafety infrastructure in low- and middle-income countries should not be treated as an afterthought, as a cooperative effort to build governance frameworks that will protect everyone.
Sources and Further Reading:
Regulating Synthetic Biology and Advanced Genetic Interventions: Biosecurity Frameworks, Gene Therapy Oversight, Stem-Cell Manufacturing Risks and Future Policy Directions. Journal of Biosafety and Biosecurity, 14 May 2026.
Doni Bloomfield et al., AI and biosecurity: The need for governance. Science, 22 Aug 2024.