Technological advances and rising military expenditures in recent years have accelerated the development of Lethal Autonomous Weapon Systems (LAWS). Though this technology is still in its infancy, it has already transformed modern warfare. LAWS, when fully evolved, will provide means for precise and independent selection and engagement of targets without exposing soldiers to battlefield dangers. A 2025 Congressional Research Service report titled Defense Primer: U.S. Policy on LAWS classifies it as “a special class of weapon systems that use sensor suites and computer algorithms to independently identify, target and employ an onboard weapon system to engage and destroy it without manual human control.” The US Department of Defense Directive 3000.09, Autonomy in Weapon Systems (2023), defined LAWS as systems that, once activated, “can select and engage targets without further intervention by a human operator.” This concept, known as “human out of the loop” or “full autonomy,” involves target selection and engagement based on inputs from artificial intelligence (AI), big data analytics, and sensor-based identification. 

According to Data M Intelligence, the global autonomous weapons market reached USD 14.2 billion in 2024 and is expected to grow to USD 33.47 billion by 2032, with a compound annual growth rate of 11.39 percent during 2025-2032. Simultaneously, global civil society initiatives are advocating a ban on fully autonomous systems. In October 2012, Amnesty International launched the Stop Killer Robots campaign, an alliance of over 180 organizations across 65 countries, calling for an international law on autonomy in weapon systems to ensure machines are not allowed to make decisions that affect life and death. 

Concerns have arisen over unsupervised use and the potential for system errors that can cause unintended civilian casualties, escalate conflicts, and threaten global peace and security. The increasing integration of autonomous weapon systems in combat has already been highlighted by their reported use in Ukraine conflict and in Gaza. A February 2025 report by the Foundation for Political, Economic and Social Research titled Deadly Algorithms: Destructive Role of Artificial Intelligence in Gaza War revealed that Israel employed AI-based systems, Lavender and Habsora, to identify and attack human targets. The report states that Lavender can approve targets within 20 seconds, often without substantive human review. Since October 2023, the system has compiled a list of 37,000 potential individuals labelled as Hamas members without verifying their military profile.  

Since 2014, the United Nations Convention on Certain Conventional Weapons (UN CCW) has debated the regulation of LAWS. In May 2024, Arms Campaign Director Steve Goose of Human Rights Watch warned that “the world is approaching a tipping point for acting on concerns over autonomous weapons systems,” underscoring the urgency of an international legal instrument. On 2 December 2024, the UN General Assembly adopted Resolution A/RES/79/62 on LAWS by 166 votes in favor, 3 against, and 15 abstentions. The resolution marked a decisive step in acknowledging global concerns over autonomous weapon systems, affirmed the applicability of international humanitarian law (IHL) and called for further consultations in 2025. The first UNGA meeting on autonomous weapons, held on 12-13 May 2025 and attended by 96 countries, including representatives from the International Committee of the Red Cross (ICRC) and civil society, reinforced momentum to prohibit and regulate LAWS. On that occasion, UN Secretary-General António Guterres advocated for a legally binding instrument to ban LAWS by 2026, describing them as “politically unacceptable and morally repugnant.”  

Despite global concerns, progress on a legally binding treaty on LAWS remains elusive due to divergent strategic interests of major powers. The US continues to resist codification of a new binding framework, emphasizing the adequacy of national weapons review mechanisms to preserve strategic and technological flexibility. While the US maintains that it does not currently possess LAWS, senior military leaders have acknowledged that Washington may be compelled to develop them if adversaries do so. Russia has opposed any binding treaty, while China supports negotiations on the CCW and the development of norms “when conditions are ripe.” The European Union, in contrast, advocates for a legally binding international instrument, emphasizing Meaningful Human Control (MHC) and compliance with IHL. The EU’s approach seeks to differentiate between systems that incorporate human oversight and those that operate without it. 

The integration of artificial intelligence into weapon systems also presents an increasing challenge to nuclear deterrence and strategic stability. For instance, during the sidelines of the Asia-Pacific Economic Cooperation (APEC) Summit in Peru in November 2024, the then US President Joe Biden and China’s President Xi Jinping jointly pledged not to integrate AI in nuclear command-and-control systems, recognizing the catastrophic risks of automation in nuclear decision-making. However, as AI rapidly improves surveillance, missile guidance and targeting systems, it is unclear whether this restraint will hold.  

The integration of AI in nuclear forces may introduce instability into deterrence dynamics by reducing decision-making time and increasing challenges caused by algorithmic bias in early warning systems, posing the threat of false nuclear alarms. Cold War history reminds us of human judgment, central to nuclear stability, and averted catastrophes. During the Cuban Missile Crisis, the B-59 submarine incident on 27 October 1962 brought the two superpowers close to nuclear exchange when a Soviet submarine commander considered launching a nuclear-tipped torpedo under the mistaken belief that hostilities had commenced. The refusal by Vasily Arkhipov to authorize the attack prevented a potential nuclear war. Similarly, Stanislav Yevgrafovich Petrov, a lieutenant colonel in the Soviet Air Defense Forces, chose to disregard a false early-warning alert indicating an incoming US nuclear strike in 1983, preventing a global nuclear disaster. Such decision-making underscores the indispensable role of human rationality in nuclear command-and-control systems.  

As LAWS presents multifaceted threats to international peace and security, states need to consider negotiating a legally binding instrument that ensures MHC over autonomy in weapon systems. Enhancing transparency, accountability, and rigorous weapons reviews are essential to prevent destabilization and ensure that technological progress does not outpace the human element in the use of force. Confidence-building measures, such as transparency in military AI, the establishment of international verification mechanisms and a moratorium on the development and deployment of LAWS, could help mitigate future dangers. 

Jawad Ali Shah is a Research Officer at the Center for International Strategic Studies Sindh (CISSS), Pakistan. He holds a BS in International Relations from the University of Sindh, Jamshoro, Pakistan. His research areas are emerging military technologies, and South Asian nuclear deterrence and strategic stability dynamics. The views are the author’s own. 

Lethal Autonomous Weapon Systems: A New Battlefield Reality was originally published on Global Security Review.

For advocates of renewed nuclear weapons testing, stop packing for the journey to the Nevada National Security Sites (NNSS). No mushroom cloud or subterranean detonation is soon to take place. Anti-nuclear protestors should also stay home.

The truth is less exciting. No real changes will happen “immediately” that “light up the sky and shake the ground.” This is not to say that the announcement had no effect. In fact, the statement was indeed monumental and incredibly significant.

Contrary to public perception, the US has never ceased testing its nuclear weapon systems. What has changed since the 1992 self-imposed moratorium on high-yield explosive testing is the nature of those tests.

Before 1992, the US conducted 1,054 nuclear weapon test explosions. The country detonated 839 of those warheads underground, mostly at the then-named Nevada Test Site, where the last halted test, Icecap, still stands as a memorial to the explosive testing days.

Several scientists involved in Icecap acknowledge that, owing to the extensive preparations undertaken, such as instrumentation, computational simulation, analysis, and test rigging, the most significant insights were gained from the limited number of unsuccessful tests. In other words, there is still great confidence in the performance and reliability of the American nuclear arsenal. It is this kind of “testing” to which President Trump’s declaration is likely referring.

Since 1992, testing has been through proxy systems that simulate a nuclear explosion’s unique energy output and then uses the results to validate physics models on advanced computer systems, known as physics-based modeling. This approach provides a way to validate the physics and predict the performance of a nuclear explosion under conditions that were never known in an underground test.

Scientists continuously conduct these tests, improving and refining them as added details are learned. They often report that scientists know much more now than possible from explosive testing.

Despite the president’s directive that testing “will begin immediately,” experts agree that resuming full-scale nuclear explosive testing is a complex and time-consuming endeavor. According to the Arms Control Association, it would take at least 36 months to prepare the Nevada Test Site for contained underground detonations.

This includes environmental assessments, infrastructure upgrades, and political approvals. This does not mean that explosive testing is impossible, but it represents a clear change in policy and a national effort to move nuclear weapons to the forefront of national strategy through an active nuclear explosive testing program.

The phrase “on an equal basis” is particularly provocative. It implies that nations like Russia and China may already be conducting nuclear explosive tests or at least advancing their capabilities in ways that challenge the spirit of the Comprehensive Nuclear Test Ban Treaty (CTBT). Either of these should sound alarms and rightly must elicit a response.

The president has chosen precisely the response as outlined in the National Institute for Deterrence Studies’ (NIDS) Dynamic Parity report, where a response matches the activities of adversaries, giving them the option to continue expanding their nuclear capabilities, knowing how America will respond, or cease and return to the table to negotiate for a more stable relationship.

The announcement of an “immediate” resumption of (explosive) testing is monumental because of its effect on deterrence. In his international policy book, The Necessity of Choice, Henry Kissinger writes that deterrence is the (mathematical) product of will and capability. Few would question that the US has a nuclear arsenal and delivery systems that can cause incredible damage and harm. However, there is growing criticism and concern that the US lacks resolve to deploy its nuclear weapons even if an existential crisis arises.

Without clear signals of resolve, adversaries may doubt American willingness to act, weakening deterrence. This declaration supports that resolve without making a direct threat to any adversary. It simply puts them on notice.

Whether President Trump’s message leads to actual detonations or remains symbolic, it marks a turning point in American nuclear policy. It also aligns with the Dynamic Parity framework advocated by Curtis McGiffin and Adam Lowther, which calls for symmetrical deterrence and strategic clarity.

President Trump is demonstrating resolve, assuring allies, and highlighting American commitment to nuclear deterrence. The path forward should prioritize modernization, transparency, and diplomacy—not a return to the destructive rituals of past decades.

James C. Petrosky, PhD, is the President and Co-founder of the National Institute for Deterrence Studies and Professor Emeritus of the Air Force Institute of Technology. Views expressed in this article are the authors own. 

“Resumption of Nuclear Testing”—Not So Fast! was originally published on Global Security Review.

The reality is that Americans live in an increasingly complex threat environment, and the credibility of the nuclear deterrent ultimately depends on America’s ability to respond to technical or strategic surprise. That includes being ready, if necessary, to conduct a nuclear test.

There are multiple and specific conditions under which the US might be compelled to resume underground nuclear testing, each involving serious strategic or technical concerns that cannot be resolved through non-explosive means as directed by the Stockpile Stewardship Program obligations.

From an American strategic perspective, it is possible that a future administration or Congress could determine that the long-standing moratorium on nuclear testing no longer serves American interests. For example, if credible intelligence revealed that an adversary, such as China or Russia, were conducting yield-producing tests, particularly to develop new capabilities or gain strategic advantage, then confidence in the existing deterrence balance could be undermined.

Adversary behavior is a factor the United States cannot ignore. China and Russia maintain active test sites and appear to be positioned to resume testing on short notice. If either nation were to conduct a low-yield test that altered the strategic balance, the US would need to respond—not necessarily by testing, but by demonstrating that it is able. Detecting and interpreting data from these tests may result in the restoration of confidence in the status quo.

Should a nuclear-armed adversary employ or threaten limited nuclear use, a carefully calibrated test could be used to demonstrate resolve, reassure allies, stabilize the situation, and deter further escalation. Such signaling would carry substantial diplomatic consequences and would only be contemplated under extraordinary circumstances.

In addition to strategic drivers that might lead to the resumption of nuclear testing, various technical issues might force America’s hand. For instance, the inability to certify the stockpile through the Stockpile Stewardship Program, such as technical issues with weapons or their components, might also be a driver to resume full-scale nuclear testing.

One such condition would be the emergence of significant doubt about the reliability or safety of an existing warhead type. As the stockpile ages, performance uncertainties can develop in critical components such as plutonium pits, high explosives, or firing systems. If these concerns cannot be resolved through laboratory experiments, modeling, or subcritical testing, a nuclear test might be required to validate performance or ensure safety margins. Such a step would follow a determination by the Nuclear Weapons Council and the national laboratories that non-testing alternatives are insufficient.

A second issue involves the development and certification of new warhead designs. While current policy emphasizes life-extension programs using legacy designs, future geopolitical or technological developments could prompt the US to pursue novel nuclear systems. For example, if the Department of Defense sought a warhead optimized for hypersonic delivery or deeply buried targets, such a design might require full-scale testing for certification—particularly if it deviates from previously tested architectures.

Lastly, there is also the possibility of future weapon development. While current policy focuses on life-extension programs, emerging mission needs may eventually require new designs. If those designs fall outside the range of previously tested systems, the US may have no choice but to test them to certify performance.

Should the United States confront a situation where confidence in warhead reliability or safety can no longer be assured through non-explosive means, or where geopolitical developments erode the credibility of deterrence, a timely and technically sound return to testing may become necessary. However, should the US resume testing for any reason, a great number of challenges will need to be met and overcome.

During the four decades of active nuclear explosive testing, the US developed a strong and thorough testing infrastructure and mindset. As nuclear explosive weapon technology evolved, so did the methods of executing tests and measuring the performance of devices. Facilities, mostly in Nevada, were built and staffed to provide an environment capable of supporting test activities and all the personnel required to perform the tests.

These tests required the expertise of scientists from multiple disciplines, engineers of various specialties, program managers, environmental control technicians, and a wide array of support staff. In addition to the technical workforce, entire teams were responsible for sustaining day-to-day life at remote test sites—providing essential services such as food, water, housing, sanitation, medical support, and logistics.

A rough estimate of the numbers of personnel required to execute an active testing program can be found in a 1981 Department of Energy Nevada Operations Office Newsletter. There were 240 federal employees, 7,100 contractors (laboratory and university personnel), and 11,300 southern Nevada support jobs. Unfortunately, just bringing together the wide variety of personnel needed to execute and support testing is only meeting an obvious challenge. A more subtle challenge is relearning how to keep any explosive test from eaking out of the ground and into the atmosphere.

A resumption of testing would still require the US to meet the obligations of two in-force international treaties; the Threshold Test Ban Treaty (TTBT), which limits the explosive yield of any test to 150 kilotons (kt), and the Limited Test Ban Treaty (LTBT), which bans all above ground and underwater tests. Compliance with the 150 kt limit on explosive yield can be easily maintained because scientists from the national laboratories can confidently ensure the magnitude of the yield will not exceed the limit.

Compliance with the obligations of the Limited Test Ban Treaty presents a different challenge. The cadre of scientific, engineering, and technical experts who would conduct the test are unlikely to have ever faced the challenges of nuclear testing—ensuring the energy and radioactive debris is “contained” in the underground environment.

The cadre of experts who last tested a nuclear weapon, almost 35 years ago, had to “learn” how to meet this unique challenge. Most likely, none of the current cadre has ever been asked to deal with such a large amount of energy released in such a small time increment. Keeping a test contained underground is a vital national interest as a leak of radioactive materials from a nuclear test would cause significant harm to the nuclear enterprise.

Any resumption of nuclear explosive testing would represent a fundamental policy shift with far-reaching implications. A return to testing would affect arms control dynamics, global nonproliferation norms, and the strategic behavior of both allies and adversaries. For these reasons, the threshold for testing remains extraordinarily high, but it is not absolute. Given the challenges facing the United States, dramatic change may come when least expected. A requirement to test a nuclear weapon for strategic or technical reasons may be a part of that change.

James Ragland is a Fellow at the National Institute for Deterrence Studies. Joel Karasikis a contractor for the Defense Nuclear Weapons School.  The views expressed are their own.

It Is Time to Test Again was originally published on Global Security Review.