On December 1, 2025, global biosecurity took center stage as world leaders, scientists, and policy experts gathered to confront a mounting challenge: the growing threat posed by biological weapons and the increasing difficulty of tracing the origins of dangerous pathogens. The urgency of the moment was underscored by India’s External Affairs Minister S. Jaishankar, who warned in New Delhi that “misuse by non-state actors is no longer a distant possibility. Bioterrorism is a serious concern that the international community has to be adequately prepared for.” His remarks, delivered at a conference marking 50 years of the Biological Weapons Convention (BWC), echoed anxieties that have only intensified since the COVID-19 pandemic swept the globe.
The struggle to determine the origins of biological threats is not new. According to The Bulletin, the 2001 anthrax letter attacks in the United States, the 2007 United Kingdom foot-and-mouth outbreak, and the recent pandemic have all exposed glaring gaps in biosecurity capabilities. Despite exhaustive investigations, the origins of these outbreaks remain shrouded in uncertainty, highlighting the limitations of current forensic science in reliably tracing biological threats back to their sources.
But why is this so difficult? As The Bulletin points out, when a novel pathogen emerges, investigators are faced with a barrage of urgent questions: Is it something familiar or entirely new? Did it arise naturally, escape from a laboratory, or result from a deliberate attack? And if there are signs of human intervention, who was responsible? With biotechnology advancing at breakneck speed, these questions are becoming more pressing—and the answers, more elusive.
Advances in technology have opened two dangerous avenues. First, as more laboratories around the world take on high-risk pathogen research, the chances of accidents involving engineered organisms increase. Second, the tools of modern biotechnology may fall into the wrong hands, enabling bad actors to design pathogens that spread faster or cause more harm than anything nature has produced. In both cases, the ability to trace threats to their origin is essential—both to hold perpetrators accountable and to mount an effective response.
Encouragingly, recent breakthroughs in machine learning and genomics are beginning to offer hope. Detecting whether a pathogen has been genetically engineered—and tracing those modifications back to their source—relies on the fact that genetic engineering techniques often leave telltale fingerprints. “There’s a plethora of techniques and choices involved in engineering a genetic sequence. These design choices leave behind genetic fingerprints that could identify the designer,” The Bulletin notes. Artificial intelligence systems, trained on thousands of engineered plasmid sequences, have achieved impressive results; in a 2021 challenge, the best system correctly identified the laboratory of origin among over 1,300 options in nearly 82 percent of cases.
Yet, real-world attribution is far messier than academic exercises. Investigators often contend with degraded samples, incomplete genetic sequences, and, sometimes, adversaries who actively try to cover their tracks. As a result, experts have outlined five key recommendations to strengthen genetic engineering detection and attribution: build comprehensive datasets of both natural and engineered pathogen sequences, train robust attribution systems for pathogens (not just plasmids), expand laboratory coverage to include those working with the most dangerous organisms, develop attribution methods that can identify covert laboratories at national or regional levels, and, crucially, prioritize transparency in how AI systems make their predictions.
Transparency is particularly vital. While AI can be eerily accurate, it often operates as a “black box,” leaving human investigators in the dark about how conclusions are reached. As The Bulletin reports, researchers at Rice University have developed PlasmidHawk, an attribution system that not only achieves near state-of-the-art performance but also reveals which DNA fragments supported its predictions. This kind of interpretability is essential for building trust in the results—especially if attribution evidence is to be used in diplomatic or legal contexts.
However, significant hurdles remain. Publicly available genomes of engineered pathogens are scarce, making it difficult to train machine learning models that can handle the full spectrum of real-world threats. The global plasmid repository Addgene hosts over 1,000 viral vectors, but experts estimate that this dataset needs to be expanded by one or two orders of magnitude to be truly effective. Moreover, the laboratories that pose the greatest risks—those operating at Biosafety Level 3 or higher—are largely missing from public repositories. Collecting data from these facilities is fraught with security concerns, and may ultimately require government oversight rather than academic stewardship.
And what about secret or covert laboratories, possibly operated by state or non-state actors? Attribution at the individual lab level may be impossible, but researchers have shown that it’s feasible to predict the country or region of origin. A 2020 MIT study, for example, managed to predict the nation of origin for engineered plasmids with over 75 percent accuracy across 33 countries. Expanding such systems could help investigators at least narrow down the possibilities in the aftermath of a biological incident.
Despite these advances, genetic engineering attribution is not a silver bullet. Some engineering techniques leave no detectable signatures. Others, like serial passage, can evolve pathogens toward desired traits without direct genetic modification. Even when engineering is detected, genomic evidence alone is unlikely to definitively identify the perpetrator. Multiple lines of evidence—geographic patterns, epidemiological data, intelligence sources—will be needed to build a convincing case. And, as The Bulletin cautions, sophisticated adversaries may attempt to plant false signatures or otherwise mislead investigators.
Against this backdrop, Minister Jaishankar’s call for a modernized BWC resonates. The convention, now in its sixth decade, lacks a compliance system, a permanent technical body, and mechanisms to keep pace with scientific advances. “These gaps must be bridged in order to strengthen confidence,” Jaishankar stated. India has proposed a National Implementation Framework that would cover high-risk agent identification, oversight of dual-use research, domestic reporting, incident management, and continuous training. Assistance during biological emergencies, he added, should be “fast, practical, and purely humanitarian.”
Jaishankar also highlighted India’s contributions to global biosecurity: “India makes 60% of the world’s vaccines. India supplies over 20% of global generic medicines, with 60% of Africa’s generics coming from India. India is home to nearly 11,000 biotech startups, up from just 50 in 2014, now it is the third-largest biotech startup ecosystem worldwide.” These capabilities, he argued, position India as a trusted partner for the Global South and a committed supporter of global biosecurity.
Ultimately, the world faces a stark choice. As Jaishankar put it, “Biology must serve peace, not advance harm. Even as science races ahead, the BWC remains the guardrail between innovation and misuse in the life sciences domain.” The next 50 years will demand concerted action—modernizing conventions, building forensic capabilities, and ensuring that all nations can detect, prevent, and respond to biological risks. In an era of heightened biological risk, the ability to reliably identify the origins of engineered pathogens could be the difference between uncertainty and accountability when the next crisis strikes.
