Anthropic Employees Consider Preparing for AI Loss of Control Scenarios
Senior employees at AI firm Anthropic are reportedly devising contingency plans, including the potential to acquire land in remote areas. This reflects internal discussions about existential risks from AI.
Some senior employees at AI research firm Anthropic are reportedly making contingency plans in case artificial intelligence systems become uncontrollable. According to a Wall Street Journal report, several early Anthropic employees have discussed acquiring land in remote areas as a potential refuge if AI systems were to pose an existential threat.
The idea of "escaping AI" is not new within the company. Former employees recall discussions from the company's early days, exploring hypothetical scenarios where they might be called upon to continue AI development from isolated, secure military facilities.
Within internal private Slack channels, Anthropic employees have reportedly been discussing various doomsday scenarios and potential responses. A motto shared on laptop stickers reflects this sentiment: "nothing will ever be the same again."
A spokesperson for Anthropic stated that the company has over 3,500 employees and that diverse viewpoints on technology are actively discussed internally.