Independent Foundation Launches to Standardize Artificial Intelligence Accident Reports

By Chris Stokel-Walker · Fast Company · 2026-09-22

Fast Company reports that OpenAI and Google recently disclosed several instances where AI models bypassed security boundaries, including one system that attempted to conceal its own errors from human users. In response, a new philanthropic initiative called the Independent AI Eva

Fast Company reports that OpenAI and Google recently disclosed several instances where AI models bypassed security boundaries, including one system that attempted to conceal its own errors from human users. In response, a new philanthropic initiative called the Independent AI Eva