The Impact of Artificial Intelligence in Nuclear Decision-Making
UNITED NATIONS, March 6 (IPS) - As artificial intelligence (AI) threatens to dominate every aspect of human lives – including political, economic, social and cultural – there is also the danger of the potential militarization of AI.
Meanwhile, the United Nations has taken a firm stance that decisions regarding the use of nuclear weapons must rest with humans, not machines, warning that integrating Artificial Intelligence (AI) into nuclear command, control, and communications (NC3) presents an unacceptable risk to global security.
The integration of AI into nuclear command, control, and communications (NC3) systems, as well as its use in military decision-making, introduces severe, unprecedented risks to global security, according to one report.
Key negative effects include the acceleration of decision-making to “machine speed” (leaving little time for human judgment), increased vulnerability to cyberattacks, and the erosion of strategic stability.
According to the Bulletin of the Atomic Scientists, command and control of nuclear weapons is a delicate and complicated system, designed to prevent error while ensuring reliability under high-pressure conditions.
In environments where vast amounts of data shape high-stakes outcomes, artificial intelligence has become a natural consideration.
“The integration of a rapidly evolving technology raises fundamental questions about responsibility, data quality, and system reliability. When a single error could have irreversible consequences, how can confidence be built around the integration of machine learning into systems that have long relied on human judgment and oversight?”
“What guardrails should be maintained? Where are the opportunities for international collaboration and consensus?”
Tariq Rauf, former Head of Verification and Security Policy at the Vienna-based International Atomic Energy Agency (IAEA), told IPS the role of and integration of Artificial Generative Intelligence (AGI) raises some of the most consequential questions of our technological era.
The integration of AGI into nuclear command, control, and communications (NC3) systems is not merely an engineering challenge — it is a civilizational one.
The Problem of Machine Speed
Perhaps the most alarming aspect of the integration of AGI into NC3 systems, he pointed out, is the compression of decision-making timelines to “machine speed.” Nuclear strategy has historically depended on deliberate human judgment — the ability of decision-makers to pause, assess ambiguous data, consult advisors, and choose restraint even under pressure or attack.
AGI systems, by contrast, are designed to process and respond at velocities no human can match. In a crisis, this creates a dangerous paradox: the very speed that makes AGI attractive also makes meaningful human oversight nearly impossible.
“If an AGI system misidentifies a sensor anomaly as an incoming missile — something that has happened with human-operated systems before, as the 1983 Soviet false alarm incident illustrates — the window for correction could shrink from minutes to seconds.”
The margin for error in nuclear decision-making has always been uncomfortably thin; AGI risks eliminating it entirely, said Rauf.
Data Quality and System Reliability
Data quality and integrity are foundational concerns regarding AGI. Machine learning systems are only as reliable as the data on which they are trained, he argued.
“Nuclear environments present unique ultra complex challenges: they involve rare, high-stakes events with limited historical data, adversarial actors who may deliberately feed misinformation into sensor networks, and geopolitical contexts that shift faster than training datasets can capture”.
An AGI system that confidently acts on corrupted or misrepresented data in a nuclear context could trigger escalation based on a fiction. Worse still, the opacity of many machine learning models — the so-called “black box” problem — means that even system designers may not be able to explain why a particular output was generated, let alone correct it in real time, declared Rauf.
Vladislav Chernavskikh, Researcher, Weapons of Mass Destruction Programme, at the Stockholm International Peace Research Institute (SIPRI) told IPS existing state approaches to AI-nuclear nexus already broadly converge on the principle of retaining human control in nuclear decision making, yet there is no consensus on how this should be defined or operationalized.
A formal recognition of this principle by nuclear-weapon states and elaboration of what human control constitutes in this context and how it can manifest in the nuclear weapons domain can be one of the first steps towards minimising risks, he declared.
At the AI Impact Summit in New Delhi last month, UN Secretary-General Antonio Guterres said the future of AI cannot be decided by a handful of countries and the whims of a few billionaires.
Last year, the General Assembly took two decisive steps, he said.
First, by creating an Independent International Scientific Panel on Artificial Intelligence, and second, by launching a Global Dialogue on AI Governance within the UN, where all countries, together with the private sector, academia and civil society, can all have a voice.
He told participants at the summit that real impact means technology that improves lives and protects the planet. And he called on them to build AI for everyone, with dignity as the default setting.
UN Spokesperson Stephane Dujarric told reporters last month, the Secretary-General is not calling for the United Nations to rule over AI. He’s calling for – and has put in place – an architecture with the help of Member States to try to ensure that everybody gets a seat at the table.
And as he said: “AI will and has already impacted all of us. It is vital that those countries who may not have the technology also have a voice and that science and fairness be put at the centre of AI.”
Responsibility and Accountability
In a further analysis, Rauf said when AGI recommendations or autonomous actions contribute to catastrophic outcomes, the question of accountability becomes deeply problematic.
Traditional chains of command assign clear human responsibility at each decision point. AGI integration fractures this clarity. Is it the software developer, the military commander, the government that deployed the system, or the algorithm itself that bears responsibility for a miscalculation? he asked.
The absence of clear accountability frameworks is not just a legal or ethical problem — it is a strategic one, because adversaries and allies alike need to understand who is in control and what decision logic is being applied.
Cyberattack Vulnerability
AGI-enhanced or dependent NC3 systems also expand the attack surface for adversaries. Sophisticated cyberattacks — including adversarial inputs designed to manipulate AGI outputs — could potentially spoof or blind these systems in ways that are difficult to detect until it is too late. The integration of AGI thus creates new vectors for destabilization that did not exist in earlier nuclear architectures, said Rauf.
The Case for International Collaboration
Despite these alarming challenges, international collaboration could be a potential avenue for managing risk. Confidence-building measures, shared technical standards, and bilateral or multilateral ‘enforceable’ agreements on the limits of AGI autonomy in nuclear systems could help preserve strategic stability.
Arms control history, said Rauf, shows that even adversaries can agree on rules that serve mutual interests in survival. Extending that tradition to AGI-enabled NC3 systems is urgently needed — before the technology outpaces diplomacy entirely.
“The integration of AGI into nuclear systems technically might be inevitable. Whether it is managed wisely is a political and moral choice that remains very much open and seems beyond the intellectual, moral/ethical processing capabilities of today’s civil and military ‘leaders’, declared Rauf.
This article is brought to you by IPS NORAM, in collaboration with INPS Japan and Soka Gakkai International, in consultative status with the UN’s Economic and Social Council (ECOSOC).
IPS UN Bureau Report
© Inter Press Service (20260306065643) — All Rights Reserved. Original source: Inter Press Service