Anthropic is hiring a weapons expert to stop Claude from starting AI war, salary around Rs 2.7 crore

Anthropic has been warning about the dangers of increasingly powerful AI and the possibility of it killing humanity. Now, the Claude maker is hiring a weapons expert to control how its AI can be used in weapons development.

Anthropic has been continuously warning about what could happen if AI gets out of control. In fact, many of its researchers, along with CEO Dario Amodei, have called for the AI industry to slow down the development of frontier models. They fear that if AI reaches advanced levels– which is AGI where it could get conscious— the intelligence machine could eventually wipe out humanity. And these are not just mere warnings. The company is serious — so serious that it is now hiring a weapons expert whose job will be to decide how far its Claude AI model should be allowed to go when it comes to weapons.

Anthropic has posted a new job opening for its Washington, D.C. office, and is looking for a Policy Design Manager, Conventional Weapons. The position sits within Anthropic's Safeguards organisation, which develops the policies, evaluations and enforcement systems that define how Claude can be used.

The person hired will be responsible for Anthropic's conventional weapons policy and will help decide where Claude should draw the line between legitimate research and potentially harmful weapons-related requests. “You will help define the line between prohibited weapons development and legitimate research and engineering work,” reads the job description.

According to Anthropic, the person taking up the role will also be expected to build threat models and evaluations to assess how its AI models could contribute to weapons development, while working with engineers to create guardrails, detection systems and enforcement tools.

How much is Anthropic paying?

And the company is paying handsomely for the job. Anthropic has listed an annual salary of $245,000 to $285,000, depending on the candidate and the internal level of the role. At the upper end, that works out to around Rs 2.7 crore a year, or roughly Rs 22.5 lakh per month before taxes. The company also says the position is eligible for visa sponsorship in some cases, although it notes that sponsorship is not guaranteed for every candidate.

Who can apply?

As for who can apply, Anthropic wants someone with practical expertise in weapons systems, preferably from a defence research lab, government agency or weapons company. A bachelor's degree or equivalent experience is required, while knowledge of AI, robotics, autonomous systems or aerospace engineering is preferred.

The timing of the job is particularly notable. It comes days after Anthropic CEO Dario Amodei called for frontier AI development to slow down, warning that the technology could advance faster than companies and governments can build the safeguards needed to control it. “We need to slow down the pace of AI development,” Amodei wrote, while arguing that this does not mean stopping AI development altogether.

According to Anthropic, this job role is about drawing a clear line between legitimate research and harmful weapons development. The person will assess how Claude could be used in weapons-related work and help create rules to prevent misuse. This includes looking at technologies such as software, robotics, sensors, guidance systems and autonomous systems, which can have both civilian and military uses. The role will cover conventional weapons, including autonomous systems that can select and engage targets without human approval.