Why a decade of doomsday warnings failed to slow the AI race

For over a decade, tech leaders and scientists have warned that superintelligent AI could threaten humanity. Despite these recurring alarms, the industry remains locked in a fierce, profit-driven race. The recent resignation of Anthropic safety researcher Jacob Coxon has reignited the debate, as he accused major firms of gambling with lives to achieve self-improving superintelligence. While CEOs like Anthropic’s Dario Amodei have publicly called for caution, critics argue that commercial pressures and the fear of competitors gaining an edge continue to undermine safety protocols. From the founding of OpenAI to the creation of Anthropic, the narrative that AI is only safe in specific hands has fueled a cycle of departures and new ventures. As companies face immense pressure to IPO, experts like Sarah Myers West suggest that meaningful security investments are being sacrificed, leaving the industry vulnerable to the very risks these doomsday warnings originally sought to prevent.