Leading AI labs have been hiring niche experts from across academic disciplines for years, often in the role of fact-checkers whose work is used to train future AI models. But Google's latest job opening is unusual even by those standards, and could shed light on the future the tech giants are preparing for humanity.
Google DeepMind is recruiting a team of “highly qualified physicists and nuclear engineers” to join the Chemical, Biological, Radiological, Nuclear and Explosives (CBRNE) team within the Responsible Development and Innovation division, Futurism reports .
According to the job description, the ideal candidate should have a PhD in a field such as plasma physics or nuclear engineering (or equivalent practical experience), as well as experience working with a U.S. national laboratory. In addition, an understanding of policies related to “arsenal management, nonproliferation, arms control, or threat reduction” is required.
Among the desired skills, Google notes:
experience in the field of AI security;
expertise in computational modeling;
experience working in “large laser institutions”;
the ability to develop risk assessments and mitigation measures that are “demonstrably unclassified and not subject to export controls through the creation of dual-use proxy models for weapons topics.”
Successful candidates are offered an annual salary ranging from $174,000 to $252,000 plus a 15% bonus.
The job itself will be challenging. As the job description states, the role involves assessing the security risks of so-called “advanced AI models” in the “nuclear domain.” Specifically, it involves “modeling support for nuclear fission, fusion, and proliferation, radiation hydrodynamics, magnetohydrodynamics, etc. in a demonstrably non-classified manner.”
Taken together, it looks as if Google's main artificial intelligence lab expects the company's tools to start issuing practical instructions for designing nuclear devices, which is why it's hiring nuclear physicists to help develop defense mechanisms.
How serious this scenario is remains to be seen, but Google is clearly demonstrating a willingness to invest heavily in addressing this problem.
