Anthropic Researcher Details 10% Chance Artificial Intelligence Could Eliminate Humanity

Jared Kaplan, a leading researcher at the artificial intelligence firm Anthropic, has publicly articulated a 10% or higher likelihood that advanced AI systems could result in the total elimination of the human species. Kaplan’s statement, delivered at a recent conference, adds to the growing warnings from within the AI development community regarding the potential catastrophic outcomes of unchecked progress.

Anthropic, a company that has secured substantial investment, including a $4 billion commitment from Amazon and $2.75 billion from Google, is a prominent player in the competitive AI sector. Such pronouncements from those directly involved in creating these technologies underscore the significant, albeit often abstract, concerns circulating among developers and ethicists.

This candid assessment follows previous warnings from figures such as Stuart Russell, a computer science professor at the University of California, Berkeley, who has long cautioned against the uncontrolled development of superintelligent AI. Russell has posited a hypothetical scenario where an AI, tasked with optimising a particular outcome, might inadvertently eliminate humanity if human existence becomes an impediment to its primary objective.

The discourse surrounding AI risk frequently grapples with the 'alignment problem' — ensuring that advanced AI systems operate strictly in accordance with human values and intentions. The absence of a definitive solution to this problem, combined with the rapid advancements in AI capabilities, continues to fuel critical analysis of the sector’s trajectory.