OpenAI Employees Long Warned of Security Lapses Amid Growth Obsession

Former employees describe an environment where the race to outpace competitors like Google and Anthropic eclipsed basic defensive measures. Staffers reportedly warned the board and senior leadership that OpenAI’s reliance on a single, expansive code repository—a "monorepo"—allowed internal users broad access to projects unrelated to their specific roles. Critics argue this structure, combined with minimal security oversight for internal research tools, created a porous environment for potential data exfiltration.

Internal tension peaked when researcher Leopold Aschenbrenner penned a 2023 memo titled "OpenAI is China's leading AI lab," warning that the company’s open culture could inadvertently facilitate foreign infiltration. These concerns gained urgency following the October arrest of Abdulrahman Alfozan, an engineer charged with planting a recording device and accessing internal files while allegedly communicating with a Saudi royal. Despite these events, the company has only recently pivoted, with executive Thibault Sottiaux reassigning a quarter of production engineers in August to bolster security infrastructure. OpenAI maintains that its investments in defensive measures are significant, noting that the recent hiring of Chief Information Security Officer Dane Stuckey and new internal security sprints signal a decisive shift in its protection strategy.