Policy Memo

Converging Technologies and Global Security Program

EDGE Project

A Critical Juncture: Global Security and the Age of Converging Technologies

At a global turning point, converging technologies demand coordinated action and strategies for CBRN nonproliferation and security

By Cindy Vestergaard • Christina McAllister • Julian Mueller-Kaler • Richard Cupitt • Robert A. Manning • Grant Turner • Kathleen Scoggin

Emerging Technology

July 30, 2025

Artificial intelligence, quantum technologies, synthetic biology, and additive manufacturing are converging to transform how we live, work, and govern. While these innovations are powering breakthroughs in healthcare, agriculture, and autonomous systems, they also carry serious risks: enabling new methods of deception, proliferation, and evasion. This paper explores how these dual-use technologies are reshaping the landscape of chemical, biological, radiological, and nuclear (CBRN) threats—and what the international community, particularly the G7-led Global Partnership, must do to stay ahead of them. At this critical juncture, coordinated action is essential to building a safer, more resilient future.

Editor’s Note: The following policy memo was originally commissioned by Global Affairs Canada for leading experts convened at Wilton Park to explore how WMD-relevant technologies (WMD-RT) can both threaten and support WMD non-proliferation and counter-terrorism efforts. Under Canada’s 2025 G7 presidency, the Global Partnership is developing a strategy to address WMD-RT. The paper was published in the Global Partnership Against WMD Newsletter [Issue No. 20, July 2025] and the GP website A Critical Juncture: Global Security and the Age of Converging Technologies, and republished in full below.

By Cindy Vestergaard, Senior Fellow and Director, Converging Technologies and Global Security Program

A global industrial transition is underway, moving from the digital automation and computing of the Third Industrial Revolution (Industry 3.0) towards the cyber-physical systems, and decentralized technologies that will dominate Industry 4.0. Unlike past industrial revolutions driven by steam, electricity, or digitization, this transition is marked by the simultaneous maturation, convergence, and dual-use applications of several key technologies, such as artificial intelligence (AI), additive manufacturing (AM), synthetic biology (Synbio), and quantum technologies (QT). Of these, AI, AM and QT are enabling technologies, driving convergence and accelerating innovation across all sectors by enhancing data processing, on-demand production, and complex problem-solving capabilities while synthetic biology is a generative technology, creating entirely new biological systems and functions. Together, these technologies are introducing novel risks – and potentially powerful mitigation tools – for chemical, biological, radiological and nuclear (CBRN) weapons non-proliferation and counterterrorism efforts, positioning international organizations, the technology community, academia, and government bodies at a critical juncture in navigating this transition.

The Dual Imperative: Civilian Innovation, Proliferation Risk

These technologies are already being used in CBRN civilian sectors. AI, for example, is currently employed for predictive maintenance and safety improvements across the nuclear material production lifecycle. In the chemical sector, BASF’s AI-driven Digital Farming Solutions delivers independent field-zone specific agronomic advice enabling farmers to produce their crops most efficiently¹ while Dow Chemical Company won the 2023 Artificial Intelligence Award for its Dow Paint Vision² solution for standardizing image processing to predict the durability of coatings and formulations. The application of AI to epidemiological datasets has birthed real-world tools to provide infectious disease surveillance early warning systems.³ Meanwhile, synthetic biology tools and techniques, many boosted by AI, have been applied in the creation of life-saving vaccines, to the development of novel therapeutics for diabetes and cancer, and to enhance the flavour of plant-based food, among other beneficial innovations .⁴ Quantum sensor technology is integral to atomic clocks for satellite navigation systems, magnetometers for oil exploration and archaeology, and various devices for medical imaging. The aerospace and automotive sectors widely use additive manufacturing processes, as does a large do-it-yourself (DIY) community made up of hobbyists and individuals without formal training.

At the same time, these technologies are challenging governance and traditional CBRN verification tools as they can be used by malicious actors to bypass existing detection systems, divert inspection efforts, or possibly develop entirely new CBRN proliferation pathways. Moreover, the rise of a “Fraud-as-a-Service” industry is democratizing sophisticated deception, offering platforms with instructions and ready-made tools that dramatically amplify the risks of tampering, data poisoning, and manipulation campaigns. Counterfeit parts have long troubled the nuclear industry, a concern which additive manufacturing exacerbates. These developments portend a shift away from the long-standing paradigm of “trust but verify” towards a more, digital-aware mindset of “verify then trust”, whereby the verification of data to confirm the reliability of technologies and behaviours is completed before assumptions on compliance are made.

Often unrecognized and occasionally ignored by those who push the envelope of technological advancement are the “dual use” risks that exist in a world where states of proliferation concern and certain terrorist organizations continue to seek weapons of mass destruction (WMD). Chemical weapons (CW) have been utilized with devastating effect on the battlefield for more than a century, and in recent decades have been used repeatedly by states and terrorists to target civilians, both indiscriminately and in targeted killings. Though biological weapons (BW) have rarely been unleashed, arsenals of the world’s deadliest diseases have been developed by both state and non-state actors and have the potential to inflict unparalleled global calamity. Nuclear weapons have not been used in conflict since 1945, but their obliterative menace continues to haunt humanity. Though each of these classes of WMD were born before or at the dawn of the computer era, new weapons of mass destruction related technologies (WMD-RT) threaten to expand their lethality, accessibility and availability and reduce accountability and detectability.

Effectively navigating this complex and shifting landscape requires understanding the specific impacts within each CBRN domain, including the significant opportunities they present. AI can enhance the analysis of declarations and facilitate bio-threat attribution. Quantum and bio-engineered sensors may offer unprecedented detection sensitivity. Other weapons WMD-RT, such as distributed ledger technology (DLT), offer novel methods for data integrity and provenance, which is particularly useful for tracking supply chains, including AI-generated data, surveillance video, or 3-D printed components, detailing who did what when along the data decision chain.

This preliminary horizon scan provides a landscape of current applications and the trends on how each of the four above-identified technologies are advancing and informing the development of adaptive strategies and the tools for today to help to mitigate their risks in the future.

Trends and Applications of Artificial Intelligence

Current Research and Applications

Artificial intelligence (AI) excels at digesting and analyzing vast volumes of data to uncover patterns and trends, a capability that is transforming CBRN civilian sectors. Predictive maintenance, anomaly detection, and automated optimization tools are now embedded in systems that monitor and adjust the performance of gas centrifuges⁵ and assess radiological data⁶ while platforms like BlueDot combine AI with expert insight to track global infectious disease threats, optimize clinical trial strategies, and improve outbreak response capabilities.⁷ AI also contributes to CBRN detection and response. For example, the U.S. Army’s Aerosol Vapor and Chemical Agent Detector (AVCAD) uses AI to autonomously detect chemical agents through mass spectrometry.⁸

Generative AI takes this further, using neural networks to not just analyze but also “create” new or derived content by identifying patterns and relationships within existing datasets. Large language models (LLMs), such as those behind ChatGPT, are similarly used by utilities and nuclear power operators. For example, Ontario Power Generation (OPG) created “ChatOPG” as a personal work assistant for staff ⁹ while Westinghouse’s Hive platform uses GenAI to optimize products, services and processes through the entire reactor lifecycle (design through operations) for its global customer base.¹⁰

AI’s role in scientific discovery is expanding as well.Automated discovery, where AI algorithms learn the rules of a system by identifying key parameters and then generating their own strategies to find optimal solutions, is still in its early stages of application in nuclear sciences, with few industrial use cases currently ready for integration into nuclear material production or feasibility research. ¹¹ In the chemical sector automated discovery is already being used, such as IBM’s RXN for Chemistry, the first remotely accessible “autonomous chemical laboratory” for forecasting results of chemical reactions,¹² reportedly slashing research timelines by as much as 30%.¹³

The Risks and Challenges

Ensuring the safety of artificial intelligence is paramount to guarantee systems operate reliably, transparently, and in alignment with human values. Despite widely shared principles emphasizing robustness, accountability, and the mitigation of bias, misinformation, and unintended outcomes, less than 3% of AI research focuses on safety,¹⁴ and both governments and private actors lack sufficient mechanisms to prevent misuse. As AI systems are becoming more powerful, their hallucinations, which generate plausible-sounding but false or fabricated outputs, are also getting worse.¹⁵ Unfortunately, at the Artificial Intelligence Action Summit in Paris in February 2025, no proposals by governments or the tech industry were made on risk thresholds or system capabilities that could pose severe harms, as was committed to at the AI Summit Seoul in 2024. These governance gaps are particularly acute in high-stakes domains like nuclear weapons systems, where the integration of opaque AI technologies could introduce strategic instability — lowering decision-making thresholds, increasing the risk of miscalculation, and making escalation more likely.

One of the most rapidly evolving risks is the accelerated rise in fraud and deception, with large language models (LLMs) at the core of a growing synthetic media ecosystem that includes deepfakes and other tools capable of producing highly convincing fake text, audio, and visuals. With access to vast datasets, fraudsters can automate more complex attacks with added speed, scale, and scope – using socially-engineered messaging that more perfectly mimics the style and tone of legitimate emails, texts, videos, or phone calls in almost any language. While most fraud attacks (80.3%) in 2023 were not as sophisticated; deepfake-related attacks surged by 3,000%.¹⁶ By 2024, digital forgeries overtook physical counterfeits, rising 244% year-over-year, with deepfakes making up 40% of biometric fraud.¹⁷ GenAI is also enabling a booming “fraud-as-a-service” market, where even low-skilled actors can conduct advanced identity fraud using tools and guides shared online. This shift from organized groups to amateurs is dramatically increasing both the scale and sophistication of attacks¹⁸ — posing a growing threat to sectors reliant on digital identity verification and Know Your Customer (KYC) compliance.¹⁹

AI’s dual-use nature dramatically expands the proliferation challenge by enabling seemingly benign civilian tools to be rapidly adapted for military or illicit use. The complexity and opacity of AI systems make their influence hard to trace and their manipulation difficult to detect, creating opportunities for malicious actors to exploit AI to evade detection, falsify nuclear safeguards data, and disrupt CBRN supply chains through cyberattacks, such as data poisoning, or by inserting unauthorized software or manipulating training datasets to skew outputs. Adversaries may also exploit AI for sophisticated counterfeiting supply chains, generating fake testing data and certification documents.²⁰ Simultaneously, well-meaning employees could unintentionally expose sensitive information when using third-party AI tools, highlighting the broad range of vulnerabilities introduced by the increasing integration of AI across various technological domains.

The chemical and biological domains illustrate these concerns with alarming clarity. On the chemical front, generative models like Dark NPS or retrosynthetic tools can accelerate the identification of novel threats and precursors.²¹ GenAI-driven platforms are also facilitating a deeper understanding of complex chemical behaviours, allowing researchers to innovate new materials leading to a surge in the discovery of novel compounds. In one experiment, a generative AI tool (MegaSyn) initially designed for pharmaceutical R&D was repurposed in an experiment to identify 40,000 toxic molecules, including nerve agents, in under six hours.²²

Agentic AI, autonomous systems capable of setting and pursuing goals, may be a decade away, with artificial general intelligence (AGI) likely further still. Yet, task-specific agents are already automating complex functions, utilizing multi-agent solutions that are achieving behaviours that mimic those of a skilled human. In the CBRN domain, these advances could accelerate threat development, obscure attribution, and strain nuclear safeguards, foreshadowing the far greater disruption AGI could bring.²³

These threats are magnified by the growing convergence of AI with other technologies, including automated laboratories, drones, additive manufacturing, nanotech miniaturization of labs, and advanced detection systems, which considerably lower the technical and logistical barriers for hostile actors. In response to these risks, “human-in-the-loop” approaches are often proposed to maintain oversight. However, there is limited discussion among regulators regarding the specifics of when, where, and at which points in decision chains human intervention is most critical. Norm-building efforts and development of standards are further complicated by the absence of a shared vocabulary.²⁴

Moreover, as commercial firms increasingly lead in AI-based threat detection and open-source intelligence, concerns around intellectual property, liability, and credibility may lead to biased prioritization of global events.²⁵ The result is a growing ecosystem of AI vulnerabilities that challenge traditional state oversight and blur the boundaries between commercial and national security capabilities.

Trends and Applications of Synthetic Biology

Current Research and Applications

Synthetic biology refers to a suite of powerful tools and techniques that enable the manipulation of living organisms and even the creation of entirely new ones. Once slow and labor-intensive, this field has rapidly accelerated due to breakthroughs in gene sequencing, editing, and data digitization — combined with the enabling power of AI, machine learning, and large language models. Even before the headlines about the recent resurrection of the dire wolf through synthetic biology,²⁶ this convergence has already reshaped medicine, food, and manufacturing. Synthetic biology was critical in the swift development of COVID-19 mRNA vaccines, which used engineered messenger RNA to direct the body’s cells to produce immune-triggering proteins—delivered in record time just 66 days after the virus’s genome was published.²⁷ Future vaccines will be able to be made more cheaply thanks to an even more recent breakthrough using bioengineered yeast to produce QS-21, a critical component of many vaccines that boosts immune response.²⁸

Similarly, the J. Craig Vinter Institute in partnership with the International Livestock Research Institute and the Friedrich Loeffler Institute are studying the potential of synthetic bio, including CRISPR-Cas, to isolate mutant strains of the African swine fever virus with the aim to reduce the time to develop attenuated vaccine candidates.²⁹ Ginko Bioworks and Novo Nordisk are also taking synbio tools and platforms and applying them to optimize pharmaceutical development³⁰ while Novo Nordisk global research group installed one of the largest biosynthetic pathways into a strain of brewer’s yeast to produce Vinblastine, a popular anti-cancer medication that was in short supply due to COVID-19 supply chain bottlenecks.³¹

Beyond healthcare, synthetic biology is transforming food and industrial chemistry. Companies like Impossible Foods use engineered yeast to produce soy leghemoglobin—a molecule that mimics the taste and texture of meat — helping plant-based foods appeal to a broader market. ³² These examples highlight synthetic biology’s growing role as a generative and converging force between biology and chemistry, enabling the bioproduction of specialty chemicals and therapeutic compounds through processes like biocatalysis and metabolic engineering, and “biopharming.”³³

Risks and Threats

Synthetic biology presents an evolving threat landscape with significant dual-use concerns for CBRN security. The ability to engineer biological systems more easily and rapidly through the application of AI and other advanced tools may lower the barrier for creating new and more harmful biological weapons, potentially accessible to a wider range of actors with reduced timeframes. Synthetic biology also blurs the lines between biological and chemical weapons, as simple genetic pathways could enable the production of high-potency molecules. Key immediate concerns include the potential to recreate known pathogenic viruses and enhance the danger of existing bacteria. A particularly challenging threat with unclear mitigation strategies is the potential to engineer microbes to produce harmful biochemicals within a living organism. While not an imminent threat, the research community recently recognized the potentially unprecedented risks of “mirror life”—a type of synthetic life in which all molecules would have reversed chirality. Mirror life could present major risks to humans, other animals, plants, and ecosystems, while offering limited benefits, and there is emerging agreement that it should not be created.³⁴

While possibilities like altering human physiology through novel pathogens or microbiome manipulation are currently less likely, they remain areas to watch as the technology advances. Critical indicators to monitor are developments that simplify the knowledge and technological hurdles associated with building novel pathogens or engineering complex biological systems.

Technological hurdles also persist in areas like the synthesis and assembly of large DNA constructs, the successful activation of engineered organisms, and the reliable engineering of complex biosynthetic pathways.³⁵ The digitization of biology, including genetic databases, AI-assisted design tools, and advanced manufacturing techniques also increasingly exposes the world of synthetic biology to cyberthreats that could compromise highly sensitive data and critical operations ranging from clinical trials to manufacturing of advanced pharmaceuticals, or even reverse the purpose of such systems for harm.³⁶ Addressing synthetic biology’s risks and threats without choking its undeniable benefits demands a nimble, innovative, and global approach to ensure appropriate and effective biosecurity guardrails and governance.

Trends and Applications of Quantum Technologies

Current Research and Applications

Thinking of quantum simply as a new tool like AI understates its significance; it is better understood as a full-stack technological paradigm shift. Quantum is a foundational platform technology—like digital computing before it—with the potential to restructure the global digital infrastructure. At its core, quantum technology leverages the unique behaviours of matter and energy at extremely small scales to solve problems and process information in entirely new ways. This enables breakthroughs in measurement, modelling, and computation that go far beyond the limits of today’s most powerful technologies. Its implications extend across secure communications, sensing and navigation, cryptographic infrastructure, and large-scale simulation, reshaping the foundation of how we compute, communicate, and sense. As such, the potential for disruptive effects—both positive and negative—should be considered systemic.

Quantum technologies (QT) encompass a broad array of emerging capabilities typically grouped into three main domains: quantum computing and simulation, quantum communication and cryptography, and quantum sensing and metrology. While there is no universally accepted taxonomy, QT includes tools such as quantum computers, simulators, quantum key distribution networks, and quantum-enhanced sensors.

Quantum technologies are becoming less research-intensive and rapidly transitioning toward commercialization. However, some advanced quantum applications, such as atomic clocks and quantum tools for measuring acceleration, gravity, and magnetic fields have been in use for decades. For example, the 1964 Mariner IV mission had the first quantum sensor in space to measure the magnetic fields of Mars, while Magnetic Resonance Imagery (MRI) epitomizes a widely used first-generation quantum technology. Today, over 500 quantum companies operate globally, including firms delivering commercial products and services in quantum sensing, quantum computing, and secure communications.³⁷ Major technology companies are investing in quantum systems—with the field spanning a growing industrial base with increasingly mature applications across multiple sectors. These capabilities are not only strategic for economic growth—they also introduce new dual-use challenges and potential risks related to non-proliferation.

Key trends include the growing use of quantum sensing for enhanced surveillance and detection of clandestine or covert activities across the nuclear, chemical, and biological threat spectrum, and quantum simulation for modeling complex biological systems or materials, such as precursors or signatures of potential WMD—which will create ever newer applications with dual relevance to both defense and civilian research. QT offers opportunities to strengthen verification regimes, detect covert activities earlier, and improve resilience against emerging nuclear, biological, and chemical threats.

Risks and Challenges

While QT is generally seen as enhancing existing capabilities rather than creating standalone weapons systems, its rapid convergence with other technologies — especially AI, space systems, and undersea capabilities — is increasingly recognized as a critical area of focus. Militaries and security alliances such as NATO³⁸ and AUKUS ³⁹ have prioritized QT as part of broader modernization strategies, particularly for their potential in ISR (intelligence, surveillance, and reconnaissance), secure communications, stealth, navigation in GPS-denied environments, precision targeting, and early threat detection. Practical applications already exist—particularly in quantum sensing and secure communications—but the most immediate wave of new applications is expected to emerge in defence contexts, including verification and validation tools that support non-proliferation objectives.

The dual-use nature of QT, coupled with limited transparency and the speed of development, continues to pose governance challenges, making it critical to develop international norms, access controls, and monitoring mechanisms that keep pace with its rapid convergence. Most studies emphasize technical hurdles, while social, governance, and cost challenges remain underexplored. QT also raises risks around strategic instability—particularly when combined with AI and advanced materials—by complicating deterrence and escalation dynamics.

Trends and Applications of Additive Manufacturing

Current Research and Applications

Traditional Additive Manufacturing (AM) involves adding material around a mandrel or similar form, contrasting with traditional Subtractive Manufacturing methods. Historically, mature technologies such as filament winding, which were used to create solid rocket casings during WWII, are now being utilized to manufacture nuclear centrifuges. Although AM is often used synonymously with 3D Printing, there are fundamental differences between the two. In contrast with traditional AM, 3D Printing, which emerged in the 1980s for rapid prototype development, involves creating objects layer by layer. The widespread use of open-source 3D technology has significantly contributed to its popularity by allowing individuals and organizations to access and modify 3D printing designs freely, fostering innovation and collaboration. In 2010, the American Society for Testing and Materials (ASTM) developed standards that classified seven 3D Printing techniques, all of which use polymers, ceramics, and metals as base materials.⁴⁰

The Organisation for the Prohibition of Chemical Weapons (OPCW) Science Advisory Board noted the key AM trend of the last two decades: that AM “has shifted from simply being a rapid prototyping technique to being incorporated in production processes.”⁴¹ Industries such as aerospace, automotive, robotics, consumer goods, energy, pharmaceuticals, and healthcare have imbued production with AM and AM-hybrid (mixing subtractive and additive manufacturing) processes. Its appeal lies in advantages like reduced waste, lower production costs, minimal need for skilled labor, and the ability to manufacture complex or obsolete parts quickly. Today, militaries worldwide are integrating AM into their supply chains and weapons systems. In Ukraine, the SPEE3D printers are deployed near the warfront, using a supersonic deposition system, where metal powder mixed with compressed air at speeds over Mach 2 bond layer by layer.⁴² Repair parts can therefore be printed in hours without wait times, lengthy procurement cycles, and with less dependence on long supply chains.

While the notion of “printing anything, anywhere” captures the imagination, real-world constraints still apply. Material limitations, specialized equipment needs, and technical requirements remain significant caveats. Creating new digital designs, using specially developed printers, and the processing of the printed items (which currently accounts for more than half of production time) to the standards required for sophisticated military or civilian applications still requires significant and varied skills. Nonetheless, these challenges haven’t slowed AM’s momentum; its untapped potential continues to fuel innovation across both civilian and military sectors. Some scholars suggest AM could broaden our understanding of weapon types, but its most immediate and tangible impact lies in enabling and enhancing existing weapons systems, particularly delivery systems, such as drones, uncrewed air systems, space launch vehicles, and uncrewed underwater equipment, as well as production of casings for traditional means of delivery, e.g., artillery shells.

AM shows promise in enabling safer handling of highly radioactive or corrosive materials, presenting both security risks and operational benefits. In the chemical and biological domains, the risk is generally considered low, with AM offering potential advantages such as proliferation-resistant production methods (e.g., chemical substitution), faster testing of bio-chem resistance, and safer handling of hazardous substances. Typically overlooked in the literature, advances in AM could also support verification regimes — enabling new sensor shapes, more durable components for inspection gear, or even rapid on-site printing of protective equipment for inspectors.

Risks and Threats

AM’s ability to produce highly customized and precise components, rapidly prototype and iterate designs, reduce production costs, and enable decentralized manufacturing increases its appeal for nefarious actors and complicates efforts to control and monitor the proliferation of WMD delivery systems. The reliance on digital design files also makes 3D printing production processes notably vulnerable to cyber risks that might attract cyber-attacks by states and non-state actors alike.

Although dual-use export controls exist, the wide range of legitimate civilian and scientific applications for AM technologies makes them challenging to define and restrict without impeding innovation or commerce. At the same time, AM poses real challenges for export control regimes, especially due to the difficulty of monitoring intangible technology transfers (ITTs), such as the cross-border flow of digital design files. The export control regimes have controls that address AM technologies, but they arise mainly as they overlap with other controlled technologies, such as high-powered lasers, and not AM technologies independently or by design. AM’s high degree of automation and reduced requirement for specialized technical labor allow for more actors—including universities, niche suppliers, and non-traditional defense partners—to participate in sensitive production processes. This diffusion of capability complicates oversight, enforcement, and verification efforts. As such, while AM presents clear benefits for innovation and industry, its integration with other disruptive technologies represents one of the most pressing and underappreciated proliferation challenges today.

Looking ahead, the convergence of AM with other emerging technologies—such as artificial intelligence, quantum sensing, and advanced materials—will likely expand its impact on both military and civilian fronts. AI-enhanced design could enable stronger, more efficient drones and underwater systems. Advances in “4D” printing may see the integration of quantum bio and chem sensors and various smart materials could protect forces in chemical or biological warfare scenarios through early detection and countermeasures. Inculcating materials with tags for identification and reporting could make monitoring and enforcement forensics easier. As innovation accelerates, many new and currently unforeseen applications are likely to emerge, further complicating the risk landscape, but creating new opportunities for building trust and assurance.

A Global Call to Action

The shift towards Industry 4.0 is ushering in an era of both unprecedented opportunity and complex risk for CBRN non-proliferation and counterterrorism. While AI, synthetic biology, AM, and QT are already revolutionizing civilian sectors, their dual-use nature demands urgent attention, as they can also empower malicious actors with sophisticated tools for deception, proliferation, and the subversion of traditional oversight. The collective challenge is to understand the nuanced impacts within each CBRN domain, moving beyond “trust but verify” to a “verify then trust” mindset, while actively collaborating to develop and implement innovative, technology-driven solutions that secure supply chains, enhance detection capabilities, and build a more resilient global security framework.

These innovations are not just transforming industries; they are reshaping how we understand and address the threats posed by chemical, biological, radiological, and nuclear (CBRN) weapons. To navigate this complex landscape effectively, we must move beyond traditional approaches and forge a collective path forward. It is a global imperative demanding coordinated action from all stakeholders:

Building on more than two decades of programming across the entire WMD threat reduction spectrum, the Global Partnership is well-placed to lead the development and implementation of technology-based WMD non-proliferation and counter-terrorism solutions to support related strategies for governments worldwide today and in the future.

This is a call to action for:

International Organizations: To update and strengthen existing WMD Conventions and instruments, develop shared vocabularies, and provide platforms for cross-sectoral dialogue on new technologies and their implications. They must act as critical conveners, leveraging their mandate to bridge gaps between governments and the private sector.
The Technology Community: To actively engage with non-proliferation experts, implement responsible AI and technology development principles, prioritize safety and security in design, and contribute to developing robust safety testing and evaluation protocols for emerging technologies. This includes investing in AI safety research, promoting ethical guidelines, and ensuring transparency regarding potential dual-use applications.
Academia: To intensify research into the dual-use implications of WMD-RT, develop comprehensive curricula that integrate ethics and security awareness for future scientists and engineers, and collaborate on independent needs assessments that can inform global policy without classification concerns.
Governments (including the G7 Global Partnership): To clearly articulate national priorities regarding WMD-RT, provide dedicated funding for research and development of non-proliferation solutions, align incentives to encourage industry engagement, and foster tailored convenings that bring together diverse stakeholders. The GP, with its two decades of experience in WMD threat reduction, is uniquely positioned to lead the development and implementation of technology-based non-proliferation and counter-terrorism solutions, promoting a “consortia of consortia” approach to accelerate innovation and deployment of dual-use technologies for global security.

The challenges are complex and the timescales tight. Yet, by embracing collaborative foresight, fostering open dialogue, and strategically leveraging technological advancements, the international community can collectively safeguard against WMD threats in this evolving era.

Notes

1
Xarvio, “Xarvio: Digitial Farming Solutions – Powered by BASF,” BASF Digital Farming GmbH, Accessed May 12, 2025: https://global.xarvio.com/.
2
DOW, “Paint Vision,” DOW Chemical Company, Accessed May 12, 2025: https://www.dow.com/paintvision/en/.
3
Bluedot, “Home,” Bluedot Global Inc., Accessed May 12, 2025: https://bluedot.global/.
4
Christopher A. Voigt, “Synthetic Biology 2020–2030: Six Commercially-Available Products That Are Changing Our World,” Nature Communications 11, no. 6379 (2020): 1-6, https://doi.org/10.1038/s41467-020-20122-2.
5
Jingjie He and Nikita Degtyarev, “AI and Atoms: How Artificial Intelligence Is Revolutionizing Nuclear Material Production,” Bulletin of the Atomic Scientists 79, no. 5 (2023): 316–28, https://doi.org/10.1080/00963402.2023.2245251.
6
Bernardo C. Bizzo, Renata R. Almeida, and Tarik K. Alkasab, “Artificial Intelligence Enabling Radiology Reporting,” Radiol Clin North Am 59, no. 6 (2021): 1045–52, https://doi.org/10.1016/j.rcl.2021.07.004.
7
Bluedot.
8
AVCAD’s development and testing phase was launched in 2019 with full rate production approval expected this year and initial operational capability by 2027. See: John Zierow, “AVCAD Advances Safety for Soldiers,” Article, United States Army, September 23, 2024, https://www.army.mil/article/279918/avcad_advances_safety_for_soldiers.
9
Ontario Power Generation. “OPG Enhances Ops with GenAI Chatbot.” (April 2024), https://genaigazette.com/opg-enhances-ops-with-genai-chatbot/
10
Westinghouse Electric Company, “Westinghouse Unveils Pioneering Nuclear Generative AI System,”Westinghouse Electric Company LLC., September 4, 2024, https://info.westinghousenuclear.com/news/westinghouse-unveils-pioneering-nuclear-generative-ai-system.
11
He and Degtyarev, “AI and Atoms.”
12
RXN for Chemistry, “Proudly Introducing the First Remotely Accessible, Autonomous Chemical Laboratory,” IBM, Accessed May 12, 2025, https://rxn.res.ibm.com/rxn/robo-rxn/welcome.
13
Team DigitalDefynd, “10 Ways AI Is Being Used in the Chemical Industry, 2025,” DigitalDefynd, Accessed May 12, 2025, https://digitaldefynd.com/IQ/ai-in-chemical-industry/.
14
Amanda McGrath and Alexandra Jonker, “What is AI Safety,” IBM, November 15, 2024, https://www.ibm.com/think/topics/ai-safety.
15
Cade Metz and Karen Weise, “A.I. Is Getting More Powerful, but Its Hallucinations Are Getting Worse,” New York Times, May 5, 2025, https://www.nytimes.com/2025/05/05/technology/ai-hallucinations-chatgpt-google.html.
16
Thomas Mcaulay, “Deepfake Fraud Attempts Are up 3000% in 2023 – Here’s Why,” The Next Web, November 15, 2023, https://thenextweb.com/news/deepfake-fraud-rise-amid-cheap-generative-ai-boom.
17
Entrust, “Identity Fraud Report 2025,” Entrust, 2024, www.entrust.com/sites/default/files/documentation/reports/2025-identity-fraud-report.pdf.
18
Ibid.
19
Ibid.
20
Sarah Case Lackner and Mara Zarka, “National Security in a Changing World: Nuclear Security and the Nuclear Supply Chain in the Age of Artificial Intelligence,” Report, Vienna Center for Disarmament and Non-Proliferation, April 17, 2025, https://vcdnp.org/report-ai-nuclear-supply-chain/.
21
Michael A. Skinnider, Fei Wang, Daniel Pasin, Russell Greiner, Leonard J. Foster, Petur W. Dalsgaard, and David S. Wishart, “A Deep Generative Model Enables Automated Structure Elucidation of Novel Psychoactive Substances,” Nature Machine Intelligence 3, (2021): 973–84, https://doi.org/10.1038/s42256-021-00407-x; Joshua Klingberg, Bethany Keen, Adam Cawley, Daniel Pasin, and Shanlin Fu, “Developments in High-Resolution Mass Spectrometric Analyses of New Psychoactive Substances,” Arch Toxicol. 96, no. 4 (2022): 949-67, https://doi.org/10.1007/s00204-022-03224-2; Xia Ning, “Generative AI for Retrosynthesis Libraries,” RePORTER, National Institutes of Health, Project Number 1R01LM014385-01, Accessed May 12, 2025, https://reporter.nih.gov/project-details/10780366.
22
Fabio Urbina, Filippa Lentzos, Cédric Invernizzi, and Sean Ekins, “Dual Use of Artificial Intelligence-Powered Drug Discovery,” Nature Machine Intelligence 4, (2022): 189–91, https://doi.org/10.1038/s42256-022-00465-9; Rebecca Sohn, “AI Drug Discovery Systems Might Be Repurposed to Make Chemical Weapons, Researchers Warn,” Scientific American, April 21, 2022, https://www.scientificamerican.com/article/ai-drug-discovery-systems-might-be-repurposed-to-make-chemical-weapons-researchers-warn/.
23
Sean Oesch, Jack Hutchins, Phillipe Austria, and Amul Chaulagain, “Agentic AI and the Cyber Arms Race,” arXiv (Computer Science), February 10, 2025, https://doi.org/10.48550/arXiv.2503.04760.
24
See for example: International Electrotechnical Commission, “Artificial
intelligence: why terminology matters,” December 2023: https://www.iec.ch/blog/artificial-intelligence-why-terminology-matters; Diligent, “AI regulations around the world: Trends, takeaways & what to watch heading into 2025”, October 2024: https://www.diligent.com/resources/guides/ai-regulations-around-the-world; UNESCO, “Enabling AI governance and innovation through standards,” https://www.unesco.org/en/articles/enabling-ai-governance-and-innovation-through-standards; and Jesús Oviedo, Moisés Rodriguez, Andrea Trenta, Dino Cannas, Domenico Natale, Mario Piattini, “ISO/IEC quality standards for AI engineering,” Computer Science Review, Vol 54, November 2024: https://doi.org/10.1016/j.cosrev.2024.100681
25
United States Government Accountability Office, “Technology Assessment: Chemical Weapons Status of Forensic Technologies and Challenges to Source Attribution,” Report to Congressional Addressees, GAO-23-105439, United States Government Accountability Office, September, 2023, https://www.gao.gov/assets/870/861287.pdf; Anthony Barrett, et al., “A Framework for Assessing and Managing Dual-Use Hazards of AI Foundation Models,” News, UC Berkeley, May 15, 2024, https://vcresearch.berkeley.edu/news/framework-assessing-and-managing-dual-use-hazards-ai-foundation-models.
26
Pandora Dewan and Patrick Pester, “Adorable Dire Wolf Pups Mark ‘World’s First De-Extinction,’ Colossal Biosciences Says,” Live Science, April 7, 2025, https://www.livescience.com/animals/extinct-species/dire-wolves-are-back-from-extinction-thanks-to-genetically-engineered-pups.
27
Elie Dolgin, “Synthetic Biology Speeds Vaccine Development,” Nature Portfolio, September 28, 2020, https://www.nature.com/articles/d42859-020-00025-4.
28
University of California – Berkeley, “Synthetic Biology Breakthrough Paves the Way for Cheaper Vaccines,” SciTechDaily, May 8, 2024, https://scitechdaily.com/synthetic-biology-breakthrough-paves-the-way-for-cheaper-vaccines/.
29
“Developing Genetic Tools to Manipulate African Swine Fever Virus and Generate Attenuated Strains” https://www.jcvi.org/research/developing-genetic-tools-manipulate-african-swine-fever-virus-and-generate-attenuated.
30
Embriette Hyde, , “Gingo Bioworks and Novo Nordisk are Shifting the SynBio-BioPharma Narrative, SynBiobeta, October 2023: Ginkgo Bioworks and Novo Nordisk are Shifting the SynBio-Biopharma Narrative – SynBioBeta.
31
Elle Caruso Fitzgerald, “ARKG Offers Exposure to Advancements in Syngthetic Biology,” Vettafi, September 2022:ARKG Offers Exposure to Advancements in Synthetic Biology | ETF Trends.
32
Voigt, 2020.
33
Jonathan B. Tucker, “The Convergence of Biology and Chemistry: Implications for Arms Control Verification,” Bulletin of the Atomic Scientists, November 1, 2010, https://thebulletin.org/2010/11/the-convergence-of-biology-and-chemistry-implications-for-arms-control-verification/.
34
Katarzyna P. Adamala, Deepa Agashe, Yasmine Belkaid et al, “Confronting Risks of Mirror Life,”
Science, Vol 386, Issue 6728, December 12, 2024: https://www.science.org/stoken/author-tokens/ST-2327/full; Adamala et al. Technical Report on Mirror Bacteria: Feasibility and Risks. December 2024. https://doi.org/10.25740/cv716pj4036, and Risks from Mirror Life, Spirit of Asilomar Entreaty 2025.4.4: https://repository.rice.edu/server/api/core/bitstreams/03831782-2fe2-4b00-82ef-781e9c8a0353/content
35
National Academies of Sciences, Engineering, and Medicine, Biodefense in the Age of Synthetic Biology (Washington, DC: The National Academies Press, 2018), https://doi.org/10.17226/24890.
36
Eleonore Pauwels, AI x Biotech: Its Cybersecurity Implications (Berlin: Konrad Adenauer Stiftung, 2025), https://www.kas.de/en/single-title/-/content/ai-x-biotech-its-cybersecurity-implications.
37
According to communication with Quantum Industry Canada on May 25, 2025.
38
NATO, “NATO releases first ever quantum strategy,” January 17, 2024: https://www.nato.int/cps/en/natohq/news_221601.htm.
39
UK Parliament, AUKUS Pillar 2: Advanced Capabilities, September 2024: https://commonslibrary.parliament.uk/research-briefings/cbp-9842/.
40
Rajeev Ranjan, Deepak Kumar, Manoj Kundu, and Subhash Chandra Moi, “A Critical Review on Classification of Materials Used in 3D Printing Process,” Materials Today: Proceedings 61, no. 1 (2022): 43-9, https://doi.org/10.1016/j.matpr.2022.03.308.
41
Director-General, “Report of the Scientific Advisory Board on Developments in Science and Technology to the Fifth Special Session of the Conference of the States Parties
to Review the Operation of the Chemical Weapons Convention,” Report, Special Session of the Conference of the States Parties to Review the Operation of the Chemical Weapons Convention, Organisation for the Prohibition of Chemical Weapons, February 22, 2023, https://www.opcw.org/sites/default/files/documents/2023/02/rc5dg01%28e%29.pdf.
42
Carolyn Schwaar, “Metal 3D Printers at Ukraine’s Frontlines Make Critical Spare Parts,” Forbes, September 20, 2023, https://www.forbes.com/sites/carolynschwaar/2023/09/20/metal-3d-printers-at-ukraines-frontlines-make-critical-spare-parts/; SPEE3D, “SPEE3D and the Australian Government: Empowering Ukraine’s Defence,” News, SPEE3D, November 14, 2023, https://www.spee3d.com/spee3d-and-the-australian-government-empowering-ukraines-defence/.

Recent & Related

Resource

Nuclear Security News and Member Updates Roundup, March 2026

Christina McAllister • Alexander Farley

Policy Paper

Securing the Future: Building the US Nuclear Security Workforce Pipeline

Christina McAllister

Field Note

Breaking Silos to Beat Scams: Why Holistic Law Enforcement Matters

Courtney Weatherby • Allison Pytlak