Anthropic Safety Researchers Resign Warning AI Labs Are Gambling With Human Existence

Prominent AI safety researcher Jacob Coxon has resigned from Anthropic - following three years of model pretraining work across both Anthropic and OpenAI- warning that leading AI labs are driving recklessly toward self-improving superintelligence. Accusing major tech executives of publicly downplaying catastrophic hazards while expressing deep anxiety in private, Coxon stated that current development trajectories represent a direct gamble with human existence. Supporting these concerns, Anthropic Alignment Science Lead Evan Hubinger publicly estimated a greater than 10% chance that frontier AI could destroy humanity within the next decade, admitting that labs currently lack a viable plan to solve alignment for superintelligent systems.High-profile resignations and explicit existential risk declarations from core alignment researchers intensify public and political scrutiny on frontier AI development. These departures strengthen calls for binding international governance frameworks, safety pause mandates, and legislative interventions to prevent unchecked capabilities scaling ahead of safety controls.As the Maldives increasingly digitizes public sector and educational infrastructure using commercial AI systems, growing unreliability or unaligned automated behaviors in frontier models could expose critical digital infrastructure to sudden vulnerabilities and systemic technological dependency risks.