Some long-serving Anthropic employees are reportedly exploring the purchase of remote land in the U.S. as a refuge in the event of an AI-related catastrophe.
Key facts
- •Some of Anthropic's longest-serving employees are considering buying remote land in the U.S. as a potential refuge.
- •The plans are reportedly a contingency measure in case AI development spirals out of control.
- •Early employees previously discussed scenarios involving relocation to shielded government facilities.
- •The report identifies ties between early Anthropic and OpenAI hires and the Effective Altruism subculture.
- •The Bay Area-based group has spent years analyzing various end-of-the-world scenarios.
Some of Anthropic's earliest employees are reportedly considering purchasing land in remote parts of the United States. According to a report by the Wall Street Journal, these staff members are making contingency plans in case artificial intelligence spirals out of control.
Historical context of AI concerns
The report notes that discussions regarding potential AI risks have occurred among employees since the company's early days. Former employees recalled company dinners where staff discussed scenarios similar to the Manhattan Project, including the possibility of relocating to electromagnetically shielded government facilities to continue AI development.
Links to Effective Altruism
The report links these concerns to the Effective Altruism (EA) subculture and a network of individuals in the Bay Area often referred to as "doomers." This group has historically focused on catastrophic risks, ranging from supervolcanoes and asteroid impacts to uncontrolled AI, a concern popularized by Eliezer Yudkowsky in the mid-2000s.
Advertisement
This article was independently rewritten by ManyPress editorial AI from reporting originally published by The Decoder.
