Chemical Disarmament in a Technologically Evolving World.

 

Statements about fast-paced technological change and the challenges this can bring to Convention are frequently made, begging the question of what should be monitored and reviewed. With much of the operational aspects of the Convention built around its verification regime and the routine declarations and inspections that the OPCW undertakes, developments related to chemical industry, the adoption and integration ofemerging technologies (and especially the rise ofartificial intelligence (AI) and automation in chemical industry, e.g., “Industry 4.0”) would certainly be relevant. In Industry 4.0, technologies and advanced data analytics (e.g., AI) are integrated into production and business processes, these technologies can include Internet of Things (IoT) devices, additive manufacturing, and automation/robotic systems. Understanding how such technologies are adopted and integrated into chemical production processes provides insight into what is required to use such systems and their advantages and limitations; the chemical industry is a global endeavor, yet industry 4.0 has not been universally adopted and implemented across the chemical sectors. Biotechnology also can produce chemicals and has seen much growth in the bioeconomy sector. The SAB has long discussed biomediated production processes, especially in the context of Discrete organic chemical (DOC) production, and its impact on the Convention, arguing that the term “produced by synthesis” (one of the criteria for whether or not a facility that produces DOC is subject to declaration) applies equally to chemicals produced by traditional chemical processes and biological methods. However, many States are of the view that “produced by synthesis” does not apply to biological processes and oversight of these types of chemical production facilities falls under policy domains outside the Convention’s purview. Additionally, regarding biomediated processes, it should be appreciated that just because a chemical process is not subject to oversight under the mandate of the Convention, it does not mean that the process is not regulated: as chemical production on its own is subject to several different oversight mechanisms, these processes, when operating at scales relevant to the Convention, would certainly be regulated under frameworks relevant to chemical safety, protection of health and environment, waste management, and others. Biotechnology processes that generate Scheduled Chemicals would, of course, be relevant, and it may be worth asking the question as to whether or not they give any advantage for making Scheduled chemicals. The Scientific Advisory Board (SAB) has suggested they may not, unless they produce naturally occurring molecules like the two Scheduled toxins). It should also be appreciated that the Scheduled chemicalwarfare agents are designed to harm life processes, such that specific biomediated production methods may also not be amenable to their presence. Regulatory and oversight frameworks for chemicals, especially for chemical security purposes, commonly rely on lists of chemicals of concern. Yet, implementation based on lists is fraught with complications and challenged by an infinite universe of known and yet-to-be-discovered chemical substances, concentration-dependent chemical hazards, and a reality that no list of chemical threats can ever be comprehensive, complete, or up-todate. This is already the case with a number of the Schedules in the Annex on Chemicals. Some of the family descriptions used in Schedules 1 and 2 have an infinite number of possible chemical structures that would meet the description of the schedule. Other schedules with family descriptions that have finite numbers of possible molecular structures can still number into the millions. There is also a widespread availability of precursors for multitudes of hazardouscompounds. While there has been great success in the development and adoption of capabilities to detect and respond to known classes of chemicals with known weaponization potential, chemicals not found on control lists or deployed as weapons in non-traditional ways do pose a risk. In this context, the potential for new technologies that allow for capabilities that fundamentally change how we think about, respond to, and mitigate the effects of weaponized chemicals provide intriguing possibilities for countering chemical weapons (if these toxicants cannot be relied on to cause harm, perhaps no one would want to use them as weapons). In light of increasing levels of complexity forrecognizing, mitigating, responding to, and counteringthe proliferation of chemical threats, innovative and enabling approaches to overcome new challenges are highly relevant. These could result from the blending of AI tools with chemical measurements and observable and presumptive signatures of chemical threat agent presence and exposure (including toxidromes). The combination of these types of data streams with AI capabilities opens up the potential for a more threat-agnostic approach to the detection and identification of chemicals, where standards-free approaches might reduce the need for libraries of analytical data or the analysis a reference chemical to conform the result of a chemical measurement. The use of new technologies in such a manner opens up intriguing possibilities for the advancing capabilities to implement the Convention. However, if such tools were to be used to generate information to inform consequential decisions, what level of validation would be needed before it could be accepted? This is an important consideration for adopting any new and seemingly beneficial technological advancement: can it be trusted? Independent of the adoption of new technologies for chemical identification, continued development of analytical methods such as those used within the OPCWDesignated Laboratories for broader types and classes oftoxic chemical threats is valuable to the Convention. When verification questions become more complex than just identifying a chemical substance, validatingand adopting chemical forensics methods provide an additional means to strengthen capabilities. From a security perspective, relevant guiding questions on technologies and scientific advances might include: 

a. Are there new types of toxic chemical threat agents that we may not immediately recognize or have limited capabilities to respond to? 

b. Are there new production routes and production equipment for toxic chemical threat agents that inspectors might not recognize? 

c. Are there new developments that might enable weaponization (delivery) of toxic chemical threats or of chemicals previously discounted as threats due to difficulties with weaponization? 

d. Are there new developments that would facilitate easier accessibility of chemical weapons technology to those who traditionally would not have had the capabilities or resources to acquire it?

 e. Are there new technologies or capabilities that can help to counter any of the concerns raised in points (a) to (d) above? 

It may not be possible to answer these questions with a high degree of certainty as new scientific developments arise. Unfortunately, uncertainty about how a new development could be used may only reinforce any of the concerns raised, which once more emphasizes the need for engaging with experts to understand the true capabilities of newscience and technology and who could provide practical guidance on how they might impact the Convention. For example, take the concerns around AI because it also brings with it the capability to design new types of highly toxic chemicals. Despite the concerns raised, computational prediction of molecular properties is not new. Rather, the AI tools we have now appear to be much better and more powerful than the ones chemists have used in the past. The practice of designing new types of toxicmolecules is also commonly used for drug discovery and development—drugs designed for medical uses are chemicals intended to have “chemical actionon life processes”, the very definition of toxic chemical under the Convention. Traditional drug development processes involve screening thousands of lead compounds before a single drug is taken forward into clinical trials and eventually released to market as an approved drug; the compounds that were eliminated in the screening may have been too toxic for safe medical use or perhaps were not toxic enough to have the necessary chemical action on life processes to be an effective medicine. Chemicals used for medicinal purposes (even highly toxic substances like nitrogen mustard when used for oncology, or fentanyl, which is considered an essential medicine by the World Health Organization for the treatment of severe pain) are not chemical weapons under the definitions provided in Article II (they are being used for “purposes not prohibited” and thus are not “chemical weapons”). Such chemicals can be quite harmful if taken in high enough doses and/or administered without proper medical supervision and instruction. The same AI tools capable of designingtoxic molecules are incredibly powerful and potent for developing new and more effective medicinal chemicals. Additionally, predicting chemical properties is only a first step in realizing a new chemical substance, the actual chemical would need to be synthesized and tested to determine if it has the properties predicted by the AI. A new chemical predicted or designed by AI does not exist in the physical world until someone makes the effort to prepare it. One should also bear in mind that if a new technology does not fundamentally change how we think of, define, or counter a chemical weapon (perhaps it just provides a “new” way to do something we already know how to do), there may not be any action that could be taken beyond keeping aware of how this new technology further develops. There are also situations where policies in seemingly unrelated sectors from the security concerns of chemical weapons unexpectedly have impact on the Convention, leading to new trends and developments. For example, chemicals restricted under the Montreal Protocol (a treaty that regulates production and consumption of ozone-depleting substances) and the Stockholm Convention (a treaty intended to protect humans and the environment from persistent organic pollutants) have led to increased demand for certain types of Schedule 2 chemicals as replacements for now restricted substances that had previously been used as flame retardants. Issues such as these illustrate that there can be convergences, commonalities, challenges, or unexpected influences between diverse and seemingly unrelated agreements that alter trends and demands for chemicals with relevance to chemical agent production in applications that fall under purposes not prohibited. Finally, when looking at chemical weapon incidents over the past decade, we have examples of the deployment of chlorine gas and sulfur mustard, chemical agents from the First World War. With chlorine, it has been reported that the nefarious actors repurposed gas cylinders and transport tanks into chemical bombs. Sulfur mustard is thought to have been first reported in scientific literature in the year 1822 by chemists studying electrophilic addition chemistry, it was nearly 100 years later, during World War 1, that this chemical was adopted for use as a chemical weapon. In 2024, despite having known this chemical for over 200 years, scientific studies have not yet determined the biomolecular mechanism by which it forms its characteristic blister injuries. Other chemical attacks of the past decade include the use of G- and V-series nerve agents (chemicals known since the 1930’s and 1950’s respectively), and in 2016 a sulfur mine was intentionally set on fire, resulting in a wide area of exposure to sulfur dioxide, a toxic gas. The burning of sulfur-containing materials to produce sulfur dioxide was actually used as a method of generating toxic fumes that were suspected to have been used in Roman times.In the Salisbury incident of 2018, a “new type of nerve agent” was used. While unscheduled at the time, the OPCW Designated Laboratories, using methods developed and validated for other chemical warfare agents, were able to readily identify it. The verified destruction of the world’s declared chemical weapons stockpiles was achieved in 2023, a milestone event in the history of chemical disarmament. The two most widely used methods to destroy the chemicals in the stockpiles were combustion and hydrolysis, chemical processes that certainly wouldn’t be seen as innovative new science. These points are important: they speak to demonstrated uses of chemical threat agents that are not relying on new and emerging technologies. They require us to be diligent about perceiving the nature of chemical threats and capabilities needed to counter them. As we assess the impact of new science, we cannot not lose sight of prior knowledge and subject matter expertise (and the potential for “low-tech” approaches and methods) to pose risks, with potentially higher likelihood of occurrence to a chemical weapon using cutting-edge science. Knowledge management and scientific developments that can help us to retain and have quick access to demonstrated key capabilities, might also be thought of as a critical area of science and technology to invest in for ensuring the Convention remains fit-for-purpose.