AI Safety Leaders Abandon Restraint Under Commercial Pressure

Anthropic rejected Pentagon contracts over surveillance concerns but simultaneously abandoned its safety-first pledge, citing competitive pressures that "destroy voluntary restraint." This marks a critical shift where even AI safety leaders are deprioritizing ethical positions to compete with OpenAI and others in the rapidly accelerating model race.