Human Factors in Verifying Warhead Dismantlement
Arms control agreements that envision the verified dismantlement of nuclear weapons require the availability of suitable technology to perform the verification. But they also depend on the good faith of the participants and a shared sense of confidence in the integrity of the verification process.
An exercise in demonstrated warhead dismantlement showed that such confidence could be easily disrupted. The exercise, sponsored by the United States and the United Kingdom in 2010 and 2011, was described in a recent paper by Los Alamos scientists. See Review of the U.S.-U.K. Warhead Monitored Dismantlement Exercise by Danielle Kristin Hauck and Iain Russell, Los Alamos National Laboratory, August 4, 2016.
Participants played the roles of the host nation, whose weapons were to be dismantled, and of the monitoring nation, whose representatives were there to verify dismantlement. Confusion and friction soon developed because “the host and monitoring parties had different expectations,” the authors reported.
“The monitoring party did not expect to justify its reasons for performing certain authentication tasks or to justify its rationale for recommending whether a piece of equipment should or should not be used in the monitoring regime. However, the host party expected to have an equal stake in authentication activities, in part because improperly handled authentication activities could result in wrongful non-verification of the treaty.”
“Attempts by the host team to be involved in the authentication activities, and requests for justifications of monitoring party decisions felt intrusive and controlling. Monitoring party rebuffs to the host team reduced the host’s confidence in the sincerity of the monitoring party for cooperative monitoring.”
What emerged is that verified dismantlement of nuclear weapons is not simply a technical problem, though it is also that.
FAS and FLI partnered to build a series of convenings and reports across the intersections of artificial intelligence (AI) with biosecurity, cybersecurity, nuclear command and control, military integration, and frontier AI governance. This project brought together leaders across these areas and created a space that was rigorous, transpartisan, and solutions-oriented to approach how we should think about how AI is rapidly changing global risks.
AI is already consequential, but its future trajectory remains contested. Policymakers should make their assumptions explicit, focus on what can be shaped rather than what can be perfectly predicted, and build institutions that can learn and respond as evidence changes.
From grassroots community impacts to global geopolitical dynamics, understanding developing data center capacities is emerging as a critical analytical challenge.
The last remaining agreement limiting U.S. and Russian nuclear weapons has now expired. For the first time since 1972, there is no treaty-bound cap on strategic nuclear weapons.